diff --git a/DBCA_original_Estonian/Qwen3-14B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-770/README.md b/DBCA_original_Estonian/Qwen3-14B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-770/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/DBCA_original_Estonian/Qwen3-14B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-770/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/DBCA_original_Estonian/Qwen3-14B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-770/adapter_config.json b/DBCA_original_Estonian/Qwen3-14B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-770/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..48296ef469e96c2852463d1edbb4991ea6f987f3
--- /dev/null
+++ b/DBCA_original_Estonian/Qwen3-14B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-770/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.024002122189250963,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "v_proj",
+ "down_proj",
+ "q_proj",
+ "k_proj",
+ "up_proj",
+ "gate_proj",
+ "o_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/DBCA_original_Estonian/Qwen3-14B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-770/chat_template.jinja b/DBCA_original_Estonian/Qwen3-14B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-770/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/DBCA_original_Estonian/Qwen3-14B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-770/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/DBCA_original_Estonian/Qwen3-14B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-770/tokenizer_config.json b/DBCA_original_Estonian/Qwen3-14B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-770/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/DBCA_original_Estonian/Qwen3-14B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-770/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/DBCA_original_Estonian/Qwen3-14B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-770/trainer_state.json b/DBCA_original_Estonian/Qwen3-14B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-770/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..dea9ced5639d3647a616899d614786e61fdac621
--- /dev/null
+++ b/DBCA_original_Estonian/Qwen3-14B-Base_original_features_structural_train_original_features_structural_test1/checkpoint-770/trainer_state.json
@@ -0,0 +1,206 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.0,
+ "eval_steps": 500,
+ "global_step": 770,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.7284933185577394,
+ "epoch": 0.12987012987012986,
+ "grad_norm": 0.6036743521690369,
+ "learning_rate": 1.542172096032852e-05,
+ "loss": 1.6924661254882813,
+ "mean_token_accuracy": 0.6527876782417298,
+ "num_tokens": 133878.0,
+ "step": 50
+ },
+ {
+ "entropy": 0.9392665743827819,
+ "epoch": 0.2597402597402597,
+ "grad_norm": 0.41290706396102905,
+ "learning_rate": 3.115817091984742e-05,
+ "loss": 0.8887335968017578,
+ "mean_token_accuracy": 0.7722584992647171,
+ "num_tokens": 265533.0,
+ "step": 100
+ },
+ {
+ "entropy": 0.7594830900430679,
+ "epoch": 0.38961038961038963,
+ "grad_norm": 0.5102436542510986,
+ "learning_rate": 4.689462087936632e-05,
+ "loss": 0.7127500915527344,
+ "mean_token_accuracy": 0.8033800584077835,
+ "num_tokens": 386480.0,
+ "step": 150
+ },
+ {
+ "entropy": 0.6976428633928299,
+ "epoch": 0.5194805194805194,
+ "grad_norm": 0.45923739671707153,
+ "learning_rate": 6.263107083888522e-05,
+ "loss": 0.653647689819336,
+ "mean_token_accuracy": 0.8134089750051499,
+ "num_tokens": 520135.0,
+ "step": 200
+ },
+ {
+ "entropy": 0.7009701311588288,
+ "epoch": 0.6493506493506493,
+ "grad_norm": 0.5048617720603943,
+ "learning_rate": 7.836752079840412e-05,
+ "loss": 0.6510067749023437,
+ "mean_token_accuracy": 0.8158668893575668,
+ "num_tokens": 640303.0,
+ "step": 250
+ },
+ {
+ "entropy": 0.6815944242477417,
+ "epoch": 0.7792207792207793,
+ "grad_norm": 0.3944048285484314,
+ "learning_rate": 9.410397075792302e-05,
+ "loss": 0.6347948837280274,
+ "mean_token_accuracy": 0.8184031039476395,
+ "num_tokens": 766027.0,
+ "step": 300
+ },
+ {
+ "entropy": 0.6565104895830154,
+ "epoch": 0.9090909090909091,
+ "grad_norm": 0.34919872879981995,
+ "learning_rate": 0.00010984042071744191,
+ "loss": 0.6166216278076172,
+ "mean_token_accuracy": 0.8238802373409271,
+ "num_tokens": 894871.0,
+ "step": 350
+ },
+ {
+ "epoch": 1.0,
+ "eval_entropy": 0.6728800308819001,
+ "eval_loss": 0.6640146970748901,
+ "eval_mean_token_accuracy": 0.8102126026955935,
+ "eval_num_tokens": 981550.0,
+ "eval_runtime": 105.0345,
+ "eval_samples_per_second": 15.776,
+ "eval_steps_per_second": 1.98,
+ "step": 385
+ },
+ {
+ "entropy": 0.6388555985689163,
+ "epoch": 1.0389610389610389,
+ "grad_norm": 0.4504299461841583,
+ "learning_rate": 0.00012116578400073899,
+ "loss": 0.5894550704956054,
+ "mean_token_accuracy": 0.8289365124702454,
+ "num_tokens": 1021041.0,
+ "step": 400
+ },
+ {
+ "entropy": 0.6055982691049576,
+ "epoch": 1.1688311688311688,
+ "grad_norm": 0.4405532479286194,
+ "learning_rate": 0.00012106869552579844,
+ "loss": 0.5607559585571289,
+ "mean_token_accuracy": 0.8378401637077332,
+ "num_tokens": 1149161.0,
+ "step": 450
+ },
+ {
+ "entropy": 0.5999097889661789,
+ "epoch": 1.2987012987012987,
+ "grad_norm": 0.3318058252334595,
+ "learning_rate": 0.00012084732888635301,
+ "loss": 0.5543725204467773,
+ "mean_token_accuracy": 0.8376027828454972,
+ "num_tokens": 1278079.0,
+ "step": 500
+ },
+ {
+ "entropy": 0.6045731788873673,
+ "epoch": 1.4285714285714286,
+ "grad_norm": 0.32081809639930725,
+ "learning_rate": 0.00012050213893540995,
+ "loss": 0.5592476272583008,
+ "mean_token_accuracy": 0.836240154504776,
+ "num_tokens": 1405326.0,
+ "step": 550
+ },
+ {
+ "entropy": 0.5909591257572174,
+ "epoch": 1.5584415584415585,
+ "grad_norm": 0.3941470682621002,
+ "learning_rate": 0.00012003383495187264,
+ "loss": 0.5478679275512696,
+ "mean_token_accuracy": 0.8397009265422821,
+ "num_tokens": 1532507.0,
+ "step": 600
+ },
+ {
+ "entropy": 0.5752921235561371,
+ "epoch": 1.6883116883116882,
+ "grad_norm": 0.300540953874588,
+ "learning_rate": 0.00011944337918314997,
+ "loss": 0.5345719146728516,
+ "mean_token_accuracy": 0.8423490756750107,
+ "num_tokens": 1662747.0,
+ "step": 650
+ },
+ {
+ "entropy": 0.5910927414894104,
+ "epoch": 1.8181818181818183,
+ "grad_norm": 0.24327129125595093,
+ "learning_rate": 0.000118731984867979,
+ "loss": 0.5490537643432617,
+ "mean_token_accuracy": 0.8396700692176818,
+ "num_tokens": 1783817.0,
+ "step": 700
+ },
+ {
+ "entropy": 0.5870773965120315,
+ "epoch": 1.948051948051948,
+ "grad_norm": 0.41987577080726624,
+ "learning_rate": 0.00011790111374352309,
+ "loss": 0.5455797958374023,
+ "mean_token_accuracy": 0.8395824021100998,
+ "num_tokens": 1914573.0,
+ "step": 750
+ },
+ {
+ "epoch": 2.0,
+ "eval_entropy": 0.6101865721150086,
+ "eval_loss": 0.6203432083129883,
+ "eval_mean_token_accuracy": 0.819772845850541,
+ "eval_num_tokens": 1963100.0,
+ "eval_runtime": 105.1641,
+ "eval_samples_per_second": 15.756,
+ "eval_steps_per_second": 1.978,
+ "step": 770
+ }
+ ],
+ "logging_steps": 50,
+ "max_steps": 3850,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 500,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 4.13724721019904e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-100/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-100/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-100/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-100/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-100/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-100/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-100/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-100/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-100/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-100/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-100/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-100/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-100/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-100/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..c005548d47fa5635ae83f4317715518710711c40
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-100/trainer_state.json
@@ -0,0 +1,139 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.24125452352231605,
+ "eval_steps": 20,
+ "global_step": 100,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 5.0561259958272e+16,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1000/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1000/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1000/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1000/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1000/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1000/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1000/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1000/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1000/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1000/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1000/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1000/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1000/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1000/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..c63adcaed5f482995bf13ace55a555c129543415
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1000/trainer_state.json
@@ -0,0 +1,1084 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.4101326899879374,
+ "eval_steps": 20,
+ "global_step": 1000,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 5.0729150094336e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1020/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1020/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1020/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1020/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1020/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1020/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1020/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1020/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1020/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1020/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1020/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1020/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1020/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1020/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..ec9d8dc48a6ae1f0cc7de999d401b59edbdf216a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1020/trainer_state.json
@@ -0,0 +1,1105 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.4583835946924006,
+ "eval_steps": 20,
+ "global_step": 1020,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 5.17605631420416e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1040/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1040/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1040/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1040/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1040/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1040/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1040/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1040/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1040/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1040/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1040/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1040/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1040/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1040/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..4bde22aadc9c4b5c404091b042acb207f7760324
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1040/trainer_state.json
@@ -0,0 +1,1126 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.5066344993968634,
+ "eval_steps": 20,
+ "global_step": 1040,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 5.27477297872896e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1060/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1060/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1060/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1060/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1060/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1060/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1060/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1060/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1060/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1060/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1060/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1060/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1060/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1060/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..069a295a6f0aeae197280e6e8f11a94f1e98e440
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1060/trainer_state.json
@@ -0,0 +1,1147 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.554885404101327,
+ "eval_steps": 20,
+ "global_step": 1060,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 5.370774678528e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1080/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1080/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1080/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1080/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1080/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1080/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1080/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1080/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1080/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1080/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1080/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1080/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1080/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1080/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..7393272ef5b7d76ad865d4f836f7eb76c16d0647
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1080/trainer_state.json
@@ -0,0 +1,1168 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.60313630880579,
+ "eval_steps": 20,
+ "global_step": 1080,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 5.47206953373696e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1100/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1100/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1100/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1100/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1100/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1100/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1100/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1100/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1100/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1100/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1100/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1100/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1100/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1100/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..c8e91f10ac1257452261515ce5cfc4d7ec8f04bb
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1100/trainer_state.json
@@ -0,0 +1,1189 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.651387213510253,
+ "eval_steps": 20,
+ "global_step": 1100,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 5.57585367650304e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1120/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1120/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1120/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1120/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1120/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1120/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1120/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1120/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1120/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1120/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1120/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1120/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1120/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1120/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..9598501903700f259e1c30cce4af1c3af71efe84
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1120/trainer_state.json
@@ -0,0 +1,1210 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.6996381182147164,
+ "eval_steps": 20,
+ "global_step": 1120,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 5.68034904428544e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1140/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1140/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1140/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1140/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1140/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1140/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1140/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1140/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1140/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1140/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1140/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1140/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1140/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1140/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..505fdfbc367dba249fa8d83dd45f01262eae18f7
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1140/trainer_state.json
@@ -0,0 +1,1231 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.7478890229191797,
+ "eval_steps": 20,
+ "global_step": 1140,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 5.77882635423744e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1160/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1160/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1160/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1160/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1160/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1160/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1160/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1160/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1160/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1160/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1160/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1160/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1160/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1160/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..cee7f76ea10492380bbec699be18c084da80ac6b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1160/trainer_state.json
@@ -0,0 +1,1252 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.796139927623643,
+ "eval_steps": 20,
+ "global_step": 1160,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 5.88810881347584e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1180/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1180/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1180/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1180/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1180/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1180/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1180/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1180/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1180/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1180/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1180/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1180/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1180/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1180/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..2c32df1bdfdc0fe9b6d3c672427b354c8e22e3ce
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1180/trainer_state.json
@@ -0,0 +1,1273 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.844390832328106,
+ "eval_steps": 20,
+ "global_step": 1180,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 5.9850952863744e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-120/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-120/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-120/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-120/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-120/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-120/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-120/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-120/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-120/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-120/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-120/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-120/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-120/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-120/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ed4355596cafda5bb9299dae23d3207842d5987
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-120/trainer_state.json
@@ -0,0 +1,160 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.28950542822677927,
+ "eval_steps": 20,
+ "global_step": 120,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 6.1751428171776e+16,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1200/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1200/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1200/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1200/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1200/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1200/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1200/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1200/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1200/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1200/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1200/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1200/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1200/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1200/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..03beec3ac1644cc5a270ba795541d7a7fae92ad2
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1200/trainer_state.json
@@ -0,0 +1,1294 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.8926417370325694,
+ "eval_steps": 20,
+ "global_step": 1200,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 6.08848278441984e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1220/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1220/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1220/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1220/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1220/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1220/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1220/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1220/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1220/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1220/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1220/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1220/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1220/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1220/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..ad23db2d10e679aca2f1393495b2b429cd5d27fe
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1220/trainer_state.json
@@ -0,0 +1,1315 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.9408926417370327,
+ "eval_steps": 20,
+ "global_step": 1220,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 6.18975712352256e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1240/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1240/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1240/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1240/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1240/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1240/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1240/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1240/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1240/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1240/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1240/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1240/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1240/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1240/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..50efe4e154983b16e86b3d2cbb5fb8fcd6beaf5d
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1240/trainer_state.json
@@ -0,0 +1,1336 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.989143546441496,
+ "eval_steps": 20,
+ "global_step": 1240,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 6.28445263122432e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1260/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1260/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1260/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1260/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1260/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1260/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1260/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1260/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1260/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1260/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1260/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1260/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1260/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1260/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..b2cbdd2c1dcde173354f95f4001f044eb88bfe3a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1260/trainer_state.json
@@ -0,0 +1,1357 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.0361881785283473,
+ "eval_steps": 20,
+ "global_step": 1260,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 6.38081421772032e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1280/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1280/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1280/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1280/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1280/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1280/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1280/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1280/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1280/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1280/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1280/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1280/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1280/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1280/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..db9f568558198d305641ead1e487f0e075f752b0
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1280/trainer_state.json
@@ -0,0 +1,1378 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.0844390832328106,
+ "eval_steps": 20,
+ "global_step": 1280,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 6.49323564121344e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1300/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1300/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1300/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1300/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1300/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1300/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1300/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1300/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1300/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1300/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1300/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1300/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1300/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1300/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..b612e9f23d1f54b7ecf151b2061932b1d6731604
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1300/trainer_state.json
@@ -0,0 +1,1399 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.132689987937274,
+ "eval_steps": 20,
+ "global_step": 1300,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 6.60286687425792e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1320/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1320/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1320/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1320/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1320/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1320/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1320/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1320/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1320/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1320/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1320/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1320/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1320/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1320/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..8d8c50ee669b59fddeccf50ead8613fb89677028
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1320/trainer_state.json
@@ -0,0 +1,1420 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.180940892641737,
+ "eval_steps": 20,
+ "global_step": 1320,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 6.7166013285504e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1340/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1340/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1340/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1340/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1340/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1340/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1340/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1340/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1340/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1340/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1340/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1340/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1340/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1340/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..d5c00c59e28f50c6e3cdd0d7549cba12ec89101d
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1340/trainer_state.json
@@ -0,0 +1,1441 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.2291917973462003,
+ "eval_steps": 20,
+ "global_step": 1340,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 6.81639850800384e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1360/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1360/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1360/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1360/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1360/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1360/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1360/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1360/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1360/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1360/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1360/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1360/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1360/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1360/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..56065f7642e71a7c73c0d5163e593350f7591009
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1360/trainer_state.json
@@ -0,0 +1,1462 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.2774427020506636,
+ "eval_steps": 20,
+ "global_step": 1360,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 6.91881491035392e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1380/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1380/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1380/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1380/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1380/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1380/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1380/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1380/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1380/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1380/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1380/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1380/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1380/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1380/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..25dec7c7cf6c410a60a5dc4c5c2cc21fd1400744
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1380/trainer_state.json
@@ -0,0 +1,1483 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.325693606755127,
+ "eval_steps": 20,
+ "global_step": 1380,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 7.01778460685568e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-140/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-140/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-140/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-140/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-140/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-140/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-140/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-140/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-140/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-140/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-140/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-140/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-140/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-140/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..fec2fe282958b430ab7120f3b1ef692923817a2c
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-140/trainer_state.json
@@ -0,0 +1,181 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.33775633293124246,
+ "eval_steps": 20,
+ "global_step": 140,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 7.325617668096e+16,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1400/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1400/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1400/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1400/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1400/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1400/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1400/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1400/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1400/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1400/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1400/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1400/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1400/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1400/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..c680ab293848184ea1a4244770c4123a5a20c26f
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1400/trainer_state.json
@@ -0,0 +1,1504 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.37394451145959,
+ "eval_steps": 20,
+ "global_step": 1400,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 7.11747236707584e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1420/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1420/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1420/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1420/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1420/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1420/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1420/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1420/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1420/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1420/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1420/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1420/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1420/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1420/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..6201c12fb4f373072680597fd8f5a6596cf31b3e
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1420/trainer_state.json
@@ -0,0 +1,1525 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.422195416164053,
+ "eval_steps": 20,
+ "global_step": 1420,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 7.21436309814528e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1440/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1440/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1440/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1440/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1440/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1440/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1440/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1440/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1440/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1440/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1440/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1440/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1440/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1440/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..31b7dba34df7db4157ff993ebf589a55875631e6
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1440/trainer_state.json
@@ -0,0 +1,1546 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.470446320868516,
+ "eval_steps": 20,
+ "global_step": 1440,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 7.316492275008e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1460/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1460/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1460/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1460/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1460/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1460/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1460/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1460/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1460/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1460/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1460/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1460/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1460/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1460/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..bc2d92d47c15ab58ded708f7d72ddac1ef77a329
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1460/trainer_state.json
@@ -0,0 +1,1567 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.5186972255729794,
+ "eval_steps": 20,
+ "global_step": 1460,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 7.42217073825024e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1480/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1480/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1480/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1480/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1480/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1480/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1480/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1480/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1480/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1480/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1480/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1480/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1480/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1480/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..e5317cf155efd2610a548e5addd86ed0295703ca
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1480/trainer_state.json
@@ -0,0 +1,1588 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.5669481302774426,
+ "eval_steps": 20,
+ "global_step": 1480,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 7.52977087677696e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1500/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1500/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1500/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1500/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1500/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1500/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1500/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1500/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1500/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1500/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1500/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1500/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1500/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1500/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..df1f088a4fd7973b324ca5bff60adb9ec61027a0
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1500/trainer_state.json
@@ -0,0 +1,1609 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.615199034981906,
+ "eval_steps": 20,
+ "global_step": 1500,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 7.62582728619264e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1520/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1520/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1520/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1520/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1520/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1520/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1520/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1520/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1520/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1520/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1520/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1520/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1520/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1520/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..d3e8c90fadb248a785c85aa6e8b3ed971095da57
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1520/trainer_state.json
@@ -0,0 +1,1630 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.663449939686369,
+ "eval_steps": 20,
+ "global_step": 1520,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 7.72663659355392e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1540/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1540/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1540/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1540/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1540/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1540/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1540/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1540/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1540/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1540/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1540/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1540/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1540/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1540/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..dfe3c4a9a9e04f042a5c6914743aebac5175bf32
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1540/trainer_state.json
@@ -0,0 +1,1651 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.7117008443908324,
+ "eval_steps": 20,
+ "global_step": 1540,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 7.833669119808e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1560/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1560/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1560/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1560/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1560/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1560/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1560/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1560/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1560/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1560/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1560/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1560/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1560/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1560/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..cdae1365243634846a6d58597a9233f60dbf362b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1560/trainer_state.json
@@ -0,0 +1,1672 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.7599517490952956,
+ "eval_steps": 20,
+ "global_step": 1560,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 7.92949301335296e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1580/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1580/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1580/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1580/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1580/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1580/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1580/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1580/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1580/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1580/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1580/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1580/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1580/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1580/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..aa9b5eba1c3c77282dccc7f0e554e09437a5f585
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1580/trainer_state.json
@@ -0,0 +1,1693 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.808202653799759,
+ "eval_steps": 20,
+ "global_step": 1580,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 8.02561097108736e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-160/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-160/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-160/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-160/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-160/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-160/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-160/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-160/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-160/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-160/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-160/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-160/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-160/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-160/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..dfa1ad2375eaa15c74c6c2508ce6d8dda91833c8
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-160/trainer_state.json
@@ -0,0 +1,202 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.38600723763570566,
+ "eval_steps": 20,
+ "global_step": 160,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 8.311416572928e+16,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1600/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1600/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1600/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1600/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1600/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1600/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1600/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1600/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1600/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1600/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1600/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1600/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1600/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1600/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..8ca7a5aecd4943400da7431972682e996a10bbfd
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1600/trainer_state.json
@@ -0,0 +1,1714 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.856453558504222,
+ "eval_steps": 20,
+ "global_step": 1600,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 8.12746659986688e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1620/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1620/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1620/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1620/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1620/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1620/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1620/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1620/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1620/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1620/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1620/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1620/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1620/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1620/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..b36c4ce91bbc7935609537be818fd2db3009e3f9
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1620/trainer_state.json
@@ -0,0 +1,1735 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.9047044632086854,
+ "eval_steps": 20,
+ "global_step": 1620,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 8.23430764246272e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1640/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1640/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1640/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1640/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1640/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1640/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1640/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1640/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1640/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1640/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1640/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1640/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1640/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1640/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..031917b48422d2a0e9c0a58446e1b466d78d902c
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1640/trainer_state.json
@@ -0,0 +1,1756 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 3.952955367913148,
+ "eval_steps": 20,
+ "global_step": 1640,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 8.33664881908992e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1660/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1660/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1660/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1660/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1660/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1660/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1660/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1660/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1660/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1660/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1660/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1660/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1660/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1660/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..6b189a64053909ce32d570d15d86c55efb0a2cbf
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1660/trainer_state.json
@@ -0,0 +1,1777 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.0,
+ "eval_steps": 20,
+ "global_step": 1660,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 8.43045871487232e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1680/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1680/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1680/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1680/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1680/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1680/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1680/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1680/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1680/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1680/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1680/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1680/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1680/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1680/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..f6e5c9afc6144150495640ef15e5569c66c842fb
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1680/trainer_state.json
@@ -0,0 +1,1798 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.048250904704463,
+ "eval_steps": 20,
+ "global_step": 1680,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 8.53397614825728e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1700/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1700/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1700/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1700/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1700/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1700/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1700/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1700/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1700/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1700/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1700/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1700/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1700/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1700/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..9083a13603e1b9b8b12e55695618e2428000c611
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1700/trainer_state.json
@@ -0,0 +1,1819 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.0965018094089265,
+ "eval_steps": 20,
+ "global_step": 1700,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 8.62847333359872e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1720/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1720/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1720/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1720/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1720/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1720/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1720/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1720/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1720/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1720/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1720/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1720/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1720/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1720/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..00aef9aa1530765833d3ae53152f0c919bb74b6c
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1720/trainer_state.json
@@ -0,0 +1,1840 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.144752714113389,
+ "eval_steps": 20,
+ "global_step": 1720,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 8.73270883070208e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1740/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1740/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1740/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1740/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1740/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1740/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1740/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1740/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1740/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1740/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1740/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1740/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1740/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1740/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..b1f09486ed336471d63ae52749fb28e16757b49a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1740/trainer_state.json
@@ -0,0 +1,1861 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.193003618817853,
+ "eval_steps": 20,
+ "global_step": 1740,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 8.8321367202432e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1760/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1760/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1760/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1760/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1760/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1760/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1760/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1760/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1760/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1760/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1760/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1760/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1760/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1760/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..4b2da8e98e457fa636b30c7e072b7468b79f6f06
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1760/trainer_state.json
@@ -0,0 +1,1882 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.241254523522316,
+ "eval_steps": 20,
+ "global_step": 1760,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 8.93257673769216e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1780/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1780/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1780/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1780/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1780/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1780/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1780/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1780/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1780/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1780/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1780/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1780/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1780/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1780/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..f3ed6d4708f82f9305eb3070116267697607d0bd
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1780/trainer_state.json
@@ -0,0 +1,1903 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.2895054282267795,
+ "eval_steps": 20,
+ "global_step": 1780,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 9.03289365850368e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-180/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-180/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-180/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-180/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-180/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-180/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-180/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-180/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-180/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-180/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-180/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-180/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-180/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-180/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..d59b78954f50bccde26cf09f1b7c3756314c61f0
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-180/trainer_state.json
@@ -0,0 +1,223 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.43425814234016885,
+ "eval_steps": 20,
+ "global_step": 180,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 9.3344180170752e+16,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1800/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1800/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1800/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1800/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1800/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1800/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1800/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1800/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1800/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1800/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1800/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1800/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1800/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1800/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..a7b3a42549c7eb148702eb994cac6a8b3146119f
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1800/trainer_state.json
@@ -0,0 +1,1924 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.337756332931242,
+ "eval_steps": 20,
+ "global_step": 1800,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 9.13988515254528e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1820/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1820/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1820/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1820/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1820/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1820/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1820/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1820/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1820/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1820/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1820/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1820/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1820/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1820/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..430b73ee7bfef6eb7f7825090a1c2bec288e6890
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1820/trainer_state.json
@@ -0,0 +1,1945 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.386007237635706,
+ "eval_steps": 20,
+ "global_step": 1820,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 9.23728878627072e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1840/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1840/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1840/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1840/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1840/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1840/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1840/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1840/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1840/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1840/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1840/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1840/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1840/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1840/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..f7e2df294c8b11c1e1a16953635e388369234141
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1840/trainer_state.json
@@ -0,0 +1,1966 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.434258142340169,
+ "eval_steps": 20,
+ "global_step": 1840,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 9.3384058352256e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1860/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1860/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1860/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1860/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1860/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1860/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1860/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1860/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1860/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1860/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1860/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1860/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1860/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1860/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..4e95b0f6966009eb946a8aef9c2d7a81b04443b1
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1860/trainer_state.json
@@ -0,0 +1,1987 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.4825090470446325,
+ "eval_steps": 20,
+ "global_step": 1860,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 9.4410547534464e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1880/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1880/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1880/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1880/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1880/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1880/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1880/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1880/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1880/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1880/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1880/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1880/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1880/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1880/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..164e18ea688bb4d002277dec587a5480ce831cf4
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1880/trainer_state.json
@@ -0,0 +1,2008 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.530759951749095,
+ "eval_steps": 20,
+ "global_step": 1880,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 9.542917220928e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1900/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1900/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1900/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1900/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1900/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1900/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1900/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1900/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1900/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1900/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1900/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1900/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1900/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1900/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..752297db7dc1dbf6581b250fc03aa66ce0d8f156
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1900/trainer_state.json
@@ -0,0 +1,2029 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.579010856453559,
+ "eval_steps": 20,
+ "global_step": 1900,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 9.63221015398656e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1920/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1920/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1920/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1920/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1920/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1920/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1920/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1920/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1920/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1920/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1920/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1920/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1920/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1920/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..ac7300473661f4fba959018b2e83f301a366bc8b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1920/trainer_state.json
@@ -0,0 +1,2050 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.627261761158022,
+ "eval_steps": 20,
+ "global_step": 1920,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 9.73311520317696e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1940/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1940/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1940/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1940/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1940/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1940/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1940/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1940/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1940/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1940/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1940/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1940/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1940/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1940/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..3431eab5dc24860b5d73ecdaba810be63d4093fe
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1940/trainer_state.json
@@ -0,0 +1,2071 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.675512665862485,
+ "eval_steps": 20,
+ "global_step": 1940,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 9.83231057684736e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1960/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1960/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1960/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1960/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1960/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1960/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1960/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1960/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1960/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1960/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1960/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1960/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1960/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1960/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..dba53d485b2a049fe8943ffac61e40953b7dfad1
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1960/trainer_state.json
@@ -0,0 +1,2092 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.723763570566948,
+ "eval_steps": 20,
+ "global_step": 1960,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 9.93308569069824e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1980/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1980/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1980/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1980/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1980/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1980/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1980/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1980/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1980/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1980/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1980/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1980/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1980/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1980/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..cbbc47ec01a9eaf9bb909313c03ce45ddfcf2baf
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-1980/trainer_state.json
@@ -0,0 +1,2113 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.772014475271411,
+ "eval_steps": 20,
+ "global_step": 1980,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.00234386225792e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-20/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-20/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-20/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-20/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-20/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-20/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-20/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-20/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-20/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-20/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-20/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-20/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-20/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-20/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..da2e5cb2022017e6d5b269276d6fe00f1a17d2d1
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-20/trainer_state.json
@@ -0,0 +1,55 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.04825090470446321,
+ "eval_steps": 20,
+ "global_step": 20,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.0192401580032e+16,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-200/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-200/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-200/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-200/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-200/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-200/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-200/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-200/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-200/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-200/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-200/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-200/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-200/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-200/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..e42436e693d1ca7a674e948c9d20fd2a522fd131
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-200/trainer_state.json
@@ -0,0 +1,244 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.4825090470446321,
+ "eval_steps": 20,
+ "global_step": 200,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.03407330281472e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2000/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2000/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2000/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2000/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2000/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2000/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2000/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2000/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2000/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2000/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2000/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2000/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2000/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2000/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..be4e4afad069684c910fcf76e652adab25e205ca
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2000/trainer_state.json
@@ -0,0 +1,2134 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.820265379975875,
+ "eval_steps": 20,
+ "global_step": 2000,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.012953424664832e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2020/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2020/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2020/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2020/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2020/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2020/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2020/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2020/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2020/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2020/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2020/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2020/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2020/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2020/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..b675f6957bedb938352a3fe6aee00b952d5bce22
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2020/trainer_state.json
@@ -0,0 +1,2155 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.868516284680338,
+ "eval_steps": 20,
+ "global_step": 2020,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.023900818954496e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2040/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2040/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2040/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2040/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2040/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2040/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2040/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2040/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2040/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2040/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2040/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2040/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2040/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2040/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..1fe0a4da88b93a405a515e41cb440ae457bc4547
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2040/trainer_state.json
@@ -0,0 +1,2176 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.916767189384801,
+ "eval_steps": 20,
+ "global_step": 2040,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.03357279530624e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2060/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2060/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2060/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2060/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2060/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2060/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2060/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2060/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2060/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2060/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2060/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2060/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2060/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2060/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..e8737cbb434a9635d17ac96ac8cce3760dee2936
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2060/trainer_state.json
@@ -0,0 +1,2197 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 4.965018094089264,
+ "eval_steps": 20,
+ "global_step": 2060,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.042724346429696e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2080/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2080/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2080/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2080/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2080/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2080/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2080/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2080/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2080/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2080/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2080/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2080/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2080/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2080/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..993fca943cfb2fcf84aabe1828e1c6902284a88e
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2080/trainer_state.json
@@ -0,0 +1,2218 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.012062726176116,
+ "eval_steps": 20,
+ "global_step": 2080,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.053239278296576e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2100/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2100/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2100/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2100/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2100/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2100/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2100/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2100/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2100/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2100/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2100/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2100/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2100/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2100/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..b4a71ea8a9fe2592ca903f56ddd29fbd620f965d
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2100/trainer_state.json
@@ -0,0 +1,2239 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.060313630880579,
+ "eval_steps": 20,
+ "global_step": 2100,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.06263428721408e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2120/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2120/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2120/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2120/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2120/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2120/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2120/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2120/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2120/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2120/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2120/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2120/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2120/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2120/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..6f16bc127ee542f8d66adb128664412ae07aeb73
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2120/trainer_state.json
@@ -0,0 +1,2260 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.108564535585042,
+ "eval_steps": 20,
+ "global_step": 2120,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.072157863730688e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2140/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2140/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2140/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2140/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2140/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2140/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2140/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2140/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2140/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2140/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2140/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2140/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2140/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2140/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..feb3a676f8a79e9179b6d30082a2838297c965b8
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2140/trainer_state.json
@@ -0,0 +1,2281 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.156815440289505,
+ "eval_steps": 20,
+ "global_step": 2140,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.081538511373824e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2160/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2160/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2160/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2160/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2160/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2160/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2160/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2160/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2160/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2160/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2160/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2160/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2160/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2160/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..7c2d5f4dae29e38bd87289bbe757b038a3674300
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2160/trainer_state.json
@@ -0,0 +1,2302 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.205066344993969,
+ "eval_steps": 20,
+ "global_step": 2160,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.09116877164288e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2180/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2180/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2180/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2180/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2180/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2180/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2180/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2180/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2180/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2180/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2180/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2180/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2180/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2180/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..85bde959c830e02076108089f254192575eca1f7
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2180/trainer_state.json
@@ -0,0 +1,2323 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.253317249698432,
+ "eval_steps": 20,
+ "global_step": 2180,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.101687379312128e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-220/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-220/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-220/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-220/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-220/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-220/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-220/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-220/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-220/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-220/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-220/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-220/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-220/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-220/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..6713adf890c41d59847d6bd99da125fef06edcb8
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-220/trainer_state.json
@@ -0,0 +1,265 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.5307599517490953,
+ "eval_steps": 20,
+ "global_step": 220,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.12699758667776e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2200/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2200/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2200/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2200/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2200/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2200/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2200/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2200/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2200/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2200/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2200/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2200/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2200/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2200/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..9388a22a33243f0115a5c0148f8973d87d9932a8
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2200/trainer_state.json
@@ -0,0 +1,2344 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.301568154402895,
+ "eval_steps": 20,
+ "global_step": 2200,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.111436632997376e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2220/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2220/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2220/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2220/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2220/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2220/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2220/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2220/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2220/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2220/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2220/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2220/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2220/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2220/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..87104414f3a083196153f1674d9a7f640ff244c3
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2220/trainer_state.json
@@ -0,0 +1,2365 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.349819059107358,
+ "eval_steps": 20,
+ "global_step": 2220,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.12135822197504e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2240/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2240/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2240/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2240/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2240/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2240/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2240/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2240/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2240/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2240/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2240/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2240/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2240/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2240/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..aadb6b5387a8db696dbe17ac661dfa979a157a91
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2240/trainer_state.json
@@ -0,0 +1,2386 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.398069963811821,
+ "eval_steps": 20,
+ "global_step": 2240,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.131719539496448e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2260/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2260/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2260/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2260/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2260/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2260/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2260/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2260/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2260/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2260/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2260/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2260/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2260/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2260/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..9e7a158ebc886ab46628c85fb939bee1e11f0402
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2260/trainer_state.json
@@ -0,0 +1,2407 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.446320868516285,
+ "eval_steps": 20,
+ "global_step": 2260,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.141223967647232e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2280/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2280/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2280/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2280/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2280/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2280/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2280/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2280/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2280/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2280/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2280/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2280/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2280/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2280/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..81f36e6463d4245be279d41994c4c969613eb2ab
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2280/trainer_state.json
@@ -0,0 +1,2428 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.4945717732207475,
+ "eval_steps": 20,
+ "global_step": 2280,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.150872692411904e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2300/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2300/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2300/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2300/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2300/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2300/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2300/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2300/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2300/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2300/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2300/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2300/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2300/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2300/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..d221339dc6169d563b75cc41c62d9819d7c1c2b1
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2300/trainer_state.json
@@ -0,0 +1,2449 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.542822677925211,
+ "eval_steps": 20,
+ "global_step": 2300,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.161259997001216e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2320/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2320/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2320/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2320/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2320/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2320/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2320/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2320/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2320/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2320/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2320/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2320/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2320/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2320/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..5501ab199a4100887890d4a5b6a111c1b6da27de
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2320/trainer_state.json
@@ -0,0 +1,2470 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.591073582629674,
+ "eval_steps": 20,
+ "global_step": 2320,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.172979480755712e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2340/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2340/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2340/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2340/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2340/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2340/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2340/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2340/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2340/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2340/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2340/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2340/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2340/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2340/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..fab33e36e62b03474b5de91414850efa7de98708
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2340/trainer_state.json
@@ -0,0 +1,2491 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.639324487334138,
+ "eval_steps": 20,
+ "global_step": 2340,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.184396010008064e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2360/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2360/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2360/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2360/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2360/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2360/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2360/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2360/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2360/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2360/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2360/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2360/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2360/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2360/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..72fb9584fca4502e2ad0a5002f35b4c468b1cf55
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2360/trainer_state.json
@@ -0,0 +1,2512 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.6875753920386005,
+ "eval_steps": 20,
+ "global_step": 2360,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.194835972603392e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2380/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2380/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2380/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2380/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2380/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2380/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2380/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2380/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2380/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2380/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2380/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2380/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2380/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2380/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..9423317777e5ed83d2e04fa92025db8845fe1fbf
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2380/trainer_state.json
@@ -0,0 +1,2533 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.735826296743064,
+ "eval_steps": 20,
+ "global_step": 2380,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.20520070947584e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-240/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-240/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-240/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-240/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-240/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-240/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-240/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-240/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-240/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-240/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-240/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-240/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-240/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-240/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..b42a10b3ee92aac9a09f746a58102d8854ecbeb0
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-240/trainer_state.json
@@ -0,0 +1,286 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.5790108564535585,
+ "eval_steps": 20,
+ "global_step": 240,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.22550225143808e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2400/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2400/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2400/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2400/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2400/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2400/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2400/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2400/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2400/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2400/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2400/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2400/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2400/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2400/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..cb76cb895aefa528e078569d13da8bc33a26814f
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2400/trainer_state.json
@@ -0,0 +1,2554 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.784077201447527,
+ "eval_steps": 20,
+ "global_step": 2400,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.215871136331264e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2420/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2420/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2420/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2420/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2420/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2420/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2420/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2420/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2420/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2420/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2420/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2420/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2420/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2420/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..5b04ba34a5fa2a6264c3beca69fefc43b8b765d4
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2420/trainer_state.json
@@ -0,0 +1,2575 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.832328106151991,
+ "eval_steps": 20,
+ "global_step": 2420,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.225269564599808e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2440/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2440/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2440/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2440/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2440/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2440/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2440/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2440/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2440/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2440/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2440/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2440/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2440/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2440/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..ab529b43e5469db3c6ed09a3e65c6d790bf8abbc
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2440/trainer_state.json
@@ -0,0 +1,2596 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.8805790108564535,
+ "eval_steps": 20,
+ "global_step": 2440,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.235377166274048e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2460/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2460/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2460/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2460/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2460/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2460/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2460/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2460/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2460/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2460/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2460/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2460/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2460/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2460/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..9c3132f8d373b4981c8e8b3b3b1803a5e78ff3fe
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2460/trainer_state.json
@@ -0,0 +1,2617 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.928829915560916,
+ "eval_steps": 20,
+ "global_step": 2460,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.245342522945024e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2480/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2480/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2480/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2480/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2480/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2480/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2480/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2480/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2480/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2480/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2480/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2480/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2480/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2480/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..ad477cef895f009a270855b9b93d25d18857f9bb
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2480/trainer_state.json
@@ -0,0 +1,2638 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 5.97708082026538,
+ "eval_steps": 20,
+ "global_step": 2480,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.255260692571648e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2500/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2500/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2500/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2500/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2500/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2500/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2500/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2500/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2500/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2500/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2500/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2500/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2500/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2500/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..b832fa79e655279c26ebe87877689c22d3fac8d5
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2500/trainer_state.json
@@ -0,0 +1,2659 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.024125452352232,
+ "eval_steps": 20,
+ "global_step": 2500,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.266128245014528e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2520/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2520/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2520/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2520/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2520/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2520/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2520/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2520/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2520/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2520/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2520/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2520/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2520/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2520/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..da68528d4aab9103844424f05832743297a21052
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2520/trainer_state.json
@@ -0,0 +1,2680 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.072376357056695,
+ "eval_steps": 20,
+ "global_step": 2520,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.27587271160832e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2540/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2540/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2540/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2540/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2540/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2540/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2540/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2540/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2540/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2540/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2540/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2540/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2540/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2540/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..c4e8afdd4177689bdbea30fb98d104fda417c96a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2540/trainer_state.json
@@ -0,0 +1,2701 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.120627261761158,
+ "eval_steps": 20,
+ "global_step": 2540,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.285468094496768e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2560/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2560/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2560/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2560/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2560/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2560/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2560/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2560/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2560/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2560/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2560/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2560/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2560/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2560/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..f5038a42e21c2291cae91685685c251500a20cd5
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2560/trainer_state.json
@@ -0,0 +1,2722 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.168878166465621,
+ "eval_steps": 20,
+ "global_step": 2560,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.294447994198016e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2580/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2580/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2580/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2580/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2580/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2580/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2580/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2580/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2580/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2580/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2580/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2580/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2580/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2580/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..83d828372835c9a8ff79bd467f0e905d13b021ac
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2580/trainer_state.json
@@ -0,0 +1,2743 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.217129071170085,
+ "eval_steps": 20,
+ "global_step": 2580,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.30466022801408e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-260/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-260/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-260/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-260/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-260/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-260/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-260/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-260/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-260/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-260/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-260/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-260/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-260/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-260/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..aa701bedd558ebca14036836a1ce1960a7bfb2df
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-260/trainer_state.json
@@ -0,0 +1,307 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.6272617611580217,
+ "eval_steps": 20,
+ "global_step": 260,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.32497117319168e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2600/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2600/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2600/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2600/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2600/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2600/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2600/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2600/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2600/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2600/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2600/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2600/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2600/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2600/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..3eae41f101aec134451dc7ba46e1d7e74af02e6d
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2600/trainer_state.json
@@ -0,0 +1,2764 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.265379975874548,
+ "eval_steps": 20,
+ "global_step": 2600,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.316223789361152e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2620/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2620/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2620/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2620/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2620/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2620/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2620/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2620/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2620/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2620/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2620/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2620/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2620/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2620/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..5baaa7a33fdaffcc85a158d70bee228b6e13656d
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2620/trainer_state.json
@@ -0,0 +1,2785 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.3136308805790105,
+ "eval_steps": 20,
+ "global_step": 2620,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.327587660607488e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2640/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2640/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2640/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2640/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2640/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2640/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2640/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2640/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2640/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2640/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2640/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2640/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2640/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2640/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..e9471840db64b96661a17d7bebdfd57632391e29
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2640/trainer_state.json
@@ -0,0 +1,2806 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.361881785283474,
+ "eval_steps": 20,
+ "global_step": 2640,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.338233468135424e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2660/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2660/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2660/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2660/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2660/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2660/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2660/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2660/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2660/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2660/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2660/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2660/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2660/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2660/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..4a96b5aec58796709d8c9ba656af35eb5d7826f1
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2660/trainer_state.json
@@ -0,0 +1,2827 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.410132689987937,
+ "eval_steps": 20,
+ "global_step": 2660,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.347335096733696e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2680/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2680/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2680/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2680/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2680/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2680/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2680/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2680/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2680/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2680/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2680/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2680/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2680/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2680/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..297fa82711ccd500d575c906f4d49477e1c2ec71
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2680/trainer_state.json
@@ -0,0 +1,2848 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.458383594692401,
+ "eval_steps": 20,
+ "global_step": 2680,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.357896788226048e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2700/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2700/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2700/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2700/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2700/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2700/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2700/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2700/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2700/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2700/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2700/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2700/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2700/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2700/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..10225e9275e76874ae24352d9be2cc0fb0ed4ebd
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2700/trainer_state.json
@@ -0,0 +1,2869 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.5066344993968634,
+ "eval_steps": 20,
+ "global_step": 2700,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.367404635727872e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2720/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2720/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2720/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2720/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2720/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2720/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2720/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2720/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2720/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2720/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2720/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2720/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2720/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2720/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..458abc04006d90d92e9ce380fcc7ade8b3a9e86e
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2720/trainer_state.json
@@ -0,0 +1,2890 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.554885404101327,
+ "eval_steps": 20,
+ "global_step": 2720,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.37645634180096e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2740/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2740/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2740/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2740/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2740/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2740/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2740/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2740/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2740/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2740/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2740/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2740/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2740/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2740/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..9a5681ed6768ef4e9eaf071961979b5dd269105a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2740/trainer_state.json
@@ -0,0 +1,2911 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.60313630880579,
+ "eval_steps": 20,
+ "global_step": 2740,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.386491453233152e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2760/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2760/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2760/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2760/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2760/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2760/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2760/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2760/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2760/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2760/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2760/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2760/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2760/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2760/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..a496bf37035d382ee805a28b6b8c71d59df43973
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2760/trainer_state.json
@@ -0,0 +1,2932 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.651387213510254,
+ "eval_steps": 20,
+ "global_step": 2760,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.396458861514752e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2780/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2780/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2780/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2780/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2780/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2780/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2780/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2780/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2780/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2780/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2780/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2780/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2780/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2780/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..0492d474fe874f9693c8df18ff6c180d4f16ebe4
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2780/trainer_state.json
@@ -0,0 +1,2953 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.699638118214716,
+ "eval_steps": 20,
+ "global_step": 2780,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.40622179260416e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-280/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-280/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-280/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-280/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-280/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-280/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-280/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-280/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-280/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-280/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-280/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-280/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-280/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-280/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..5e77bca6ceb002974dd4b03f5e0de212764641b4
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-280/trainer_state.json
@@ -0,0 +1,328 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.6755126658624849,
+ "eval_steps": 20,
+ "global_step": 280,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.43159337732096e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2800/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2800/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2800/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2800/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2800/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2800/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2800/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2800/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2800/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2800/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2800/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2800/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2800/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2800/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..90069387a2330537a507e6d4bd9c264f766b64c3
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2800/trainer_state.json
@@ -0,0 +1,2974 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.74788902291918,
+ "eval_steps": 20,
+ "global_step": 2800,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.416029175257088e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2820/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2820/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2820/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2820/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2820/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2820/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2820/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2820/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2820/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2820/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2820/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2820/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2820/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2820/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..2cfbe587b395b15efebcc3eb106624fc3a147bd2
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2820/trainer_state.json
@@ -0,0 +1,2995 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.796139927623643,
+ "eval_steps": 20,
+ "global_step": 2820,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.426538892613632e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2840/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2840/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2840/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2840/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2840/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2840/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2840/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2840/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2840/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2840/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2840/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2840/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2840/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2840/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..1c5a0378e1227b966e9d1aa21abd01292f6e0c34
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2840/trainer_state.json
@@ -0,0 +1,3016 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.844390832328106,
+ "eval_steps": 20,
+ "global_step": 2840,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.436608197556224e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2860/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2860/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2860/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2860/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2860/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2860/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2860/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2860/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2860/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2860/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2860/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2860/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2860/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2860/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..5cdb5a282621a6afa6277cde7433c75fbe5faa6e
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2860/trainer_state.json
@@ -0,0 +1,3037 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.892641737032569,
+ "eval_steps": 20,
+ "global_step": 2860,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.446390277011456e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2880/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2880/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2880/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2880/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2880/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2880/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2880/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2880/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2880/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2880/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2880/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2880/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2880/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2880/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..9dac7d89b488308ea62fb72c7e67fcff1dad7009
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2880/trainer_state.json
@@ -0,0 +1,3058 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.940892641737032,
+ "eval_steps": 20,
+ "global_step": 2880,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.456912987901952e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2900/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2900/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2900/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2900/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2900/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2900/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2900/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2900/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2900/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2900/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2900/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2900/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2900/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2900/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..636dfdf808a628760e6783aae29b22ed9268428c
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2900/trainer_state.json
@@ -0,0 +1,3079 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 6.989143546441496,
+ "eval_steps": 20,
+ "global_step": 2900,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.46702811214848e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2920/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2920/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2920/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2920/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2920/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2920/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2920/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2920/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2920/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2920/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2920/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2920/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2920/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2920/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..d72c9ea492ac63726f9722d1c08b0271f0484c99
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2920/trainer_state.json
@@ -0,0 +1,3100 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.036188178528348,
+ "eval_steps": 20,
+ "global_step": 2920,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.477191363760896e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2940/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2940/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2940/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2940/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2940/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2940/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2940/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2940/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2940/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2940/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2940/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2940/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2940/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2940/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..0174205ebde1619a89bb0af92a1c3709344c0d33
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2940/trainer_state.json
@@ -0,0 +1,3121 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.084439083232811,
+ "eval_steps": 20,
+ "global_step": 2940,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.486946088407808e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2960/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2960/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2960/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2960/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2960/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2960/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2960/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2960/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2960/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2960/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2960/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2960/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2960/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2960/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..dc73b533d9bb7b4a2fd90083da3be1067b9bb65a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2960/trainer_state.json
@@ -0,0 +1,3142 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.132689987937274,
+ "eval_steps": 20,
+ "global_step": 2960,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.49675210332032e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2980/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2980/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2980/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2980/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2980/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2980/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2980/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2980/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2980/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2980/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2980/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2980/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2980/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2980/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..d0ca9a6e49dfa6b55b552d93dd0e4992d94914ea
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-2980/trainer_state.json
@@ -0,0 +1,3163 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.180940892641737,
+ "eval_steps": 20,
+ "global_step": 2980,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.5068460275904e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-300/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-300/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-300/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-300/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-300/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-300/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-300/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-300/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-300/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-300/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-300/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-300/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-300/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-300/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..9ad37206b396924887e46cfc586d8c00256cb978
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-300/trainer_state.json
@@ -0,0 +1,349 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.7237635705669482,
+ "eval_steps": 20,
+ "global_step": 300,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.53407816669184e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3000/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3000/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3000/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3000/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3000/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3000/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3000/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3000/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3000/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3000/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3000/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3000/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3000/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3000/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..1240d623acb014787904833bed3ee040106d0975
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3000/trainer_state.json
@@ -0,0 +1,3184 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.2291917973462,
+ "eval_steps": 20,
+ "global_step": 3000,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.517889847579392e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3020/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3020/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3020/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3020/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3020/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3020/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3020/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3020/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3020/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3020/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3020/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3020/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3020/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3020/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..96af121fbc179b527a06fa1b3974615f72afb375
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3020/trainer_state.json
@@ -0,0 +1,3205 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.277442702050664,
+ "eval_steps": 20,
+ "global_step": 3020,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.528316816640768e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3040/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3040/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3040/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3040/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3040/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3040/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3040/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3040/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3040/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3040/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3040/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3040/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3040/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3040/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..3c47a8c711bcdd6f96599fb94fd98839872f4873
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3040/trainer_state.json
@@ -0,0 +1,3226 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.325693606755126,
+ "eval_steps": 20,
+ "global_step": 3040,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.538312947471104e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3060/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3060/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3060/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3060/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3060/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3060/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3060/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3060/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3060/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3060/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3060/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3060/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3060/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3060/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..2ea8b7bfb69292a82e0b9d9c40da2d6434817e98
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3060/trainer_state.json
@@ -0,0 +1,3247 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.37394451145959,
+ "eval_steps": 20,
+ "global_step": 3060,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.548778213264128e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3080/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3080/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3080/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3080/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3080/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3080/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3080/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3080/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3080/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3080/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3080/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3080/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3080/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3080/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..06c52512b8b47c33e12ba4c1da1b6e9d6f5f0dd5
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3080/trainer_state.json
@@ -0,0 +1,3268 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.422195416164053,
+ "eval_steps": 20,
+ "global_step": 3080,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.55879622794112e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3100/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3100/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3100/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3100/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3100/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3100/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3100/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3100/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3100/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3100/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3100/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3100/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3100/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3100/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..ab244f842c46d639bb11bcb0cd5aba0fc61546f5
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3100/trainer_state.json
@@ -0,0 +1,3289 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.470446320868517,
+ "eval_steps": 20,
+ "global_step": 3100,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.56810028212096e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3120/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3120/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3120/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3120/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3120/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3120/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3120/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3120/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3120/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3120/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3120/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3120/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3120/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3120/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..5310dab21cf054be6f690820f92e4d8500f3f0f5
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3120/trainer_state.json
@@ -0,0 +1,3310 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.518697225572979,
+ "eval_steps": 20,
+ "global_step": 3120,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.578918424941312e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3140/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3140/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3140/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3140/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3140/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3140/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3140/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3140/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3140/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3140/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3140/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3140/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3140/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3140/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..67b4f013be7d51c89b7d212f2053efea27a66a55
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3140/trainer_state.json
@@ -0,0 +1,3331 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.566948130277443,
+ "eval_steps": 20,
+ "global_step": 3140,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.589942412694272e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3160/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3160/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3160/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3160/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3160/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3160/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3160/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3160/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3160/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3160/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3160/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3160/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3160/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3160/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..15c875b08d3a9d70515d067d1b3e603b75105423
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3160/trainer_state.json
@@ -0,0 +1,3352 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.615199034981906,
+ "eval_steps": 20,
+ "global_step": 3160,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.60040973009792e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3180/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3180/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3180/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3180/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3180/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3180/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3180/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3180/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3180/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3180/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3180/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3180/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3180/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3180/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..b9c1002b2541a7b4ce26e308422e7ba1b6f40d1b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3180/trainer_state.json
@@ -0,0 +1,3373 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.66344993968637,
+ "eval_steps": 20,
+ "global_step": 3180,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.610287551382272e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-320/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-320/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-320/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-320/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-320/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-320/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-320/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-320/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-320/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-320/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-320/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-320/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-320/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-320/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..b9e430f567f10c437447b566d58147453a96a331
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-320/trainer_state.json
@@ -0,0 +1,370 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.7720144752714113,
+ "eval_steps": 20,
+ "global_step": 320,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.64476939855872e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3200/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3200/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3200/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3200/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3200/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3200/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3200/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3200/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3200/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3200/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3200/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3200/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3200/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3200/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..9246be3490e9330967b5a46f0fb48e79e3948b74
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3200/trainer_state.json
@@ -0,0 +1,3394 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.711700844390832,
+ "eval_steps": 20,
+ "global_step": 3200,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.619632637774592e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3220/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3220/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3220/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3220/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3220/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3220/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3220/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3220/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3220/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3220/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3220/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3220/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3220/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3220/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..23bfdd4af5cfec5f6272a8cc53b4373aa32b479f
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3220/trainer_state.json
@@ -0,0 +1,3415 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.759951749095295,
+ "eval_steps": 20,
+ "global_step": 3220,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.629993271425792e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3240/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3240/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3240/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3240/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3240/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3240/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3240/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3240/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3240/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3240/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3240/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3240/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3240/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3240/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..6e47bb69c1d808ee7bdbd58c29a561f306599492
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3240/trainer_state.json
@@ -0,0 +1,3436 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.808202653799759,
+ "eval_steps": 20,
+ "global_step": 3240,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.640726614340352e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3260/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3260/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3260/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3260/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3260/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3260/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3260/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3260/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3260/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3260/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3260/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3260/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3260/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3260/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..b516e62a25158b000afee13277908fcdec2e3dae
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3260/trainer_state.json
@@ -0,0 +1,3457 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.856453558504222,
+ "eval_steps": 20,
+ "global_step": 3260,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.651297880015616e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3280/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3280/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3280/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3280/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3280/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3280/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3280/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3280/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3280/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3280/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3280/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3280/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3280/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3280/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..3834454cc26deb448425da120bb051f1256be241
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3280/trainer_state.json
@@ -0,0 +1,3478 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.904704463208685,
+ "eval_steps": 20,
+ "global_step": 3280,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.66104850144128e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3300/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3300/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3300/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3300/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3300/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3300/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3300/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3300/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3300/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3300/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3300/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3300/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3300/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3300/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..4657e289d20b4480b6e00642c82660759b4c6b8a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3300/trainer_state.json
@@ -0,0 +1,3499 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 7.952955367913148,
+ "eval_steps": 20,
+ "global_step": 3300,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.671065832248064e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3320/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3320/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3320/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3320/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3320/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3320/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3320/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3320/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3320/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3320/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3320/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3320/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3320/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3320/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..c82844a22450fd548551e3677a2f5e2c03111088
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3320/trainer_state.json
@@ -0,0 +1,3520 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.0,
+ "eval_steps": 20,
+ "global_step": 3320,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.680218836595712e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3340/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3340/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3340/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3340/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3340/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3340/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3340/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3340/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3340/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3340/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3340/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3340/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3340/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3340/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..cf8598b5de591bd79bb722d7ab5dffdf9986e3e6
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3340/trainer_state.json
@@ -0,0 +1,3541 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.048250904704464,
+ "eval_steps": 20,
+ "global_step": 3340,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.689662400297984e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3360/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3360/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3360/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3360/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3360/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3360/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3360/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3360/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3360/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3360/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3360/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3360/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3360/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3360/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..9c1516270e2abe1b6e3111eea1b9710d7def390d
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3360/trainer_state.json
@@ -0,0 +1,3562 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.096501809408926,
+ "eval_steps": 20,
+ "global_step": 3360,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.69890627389952e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3380/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3380/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3380/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3380/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3380/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3380/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3380/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3380/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3380/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3380/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3380/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3380/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3380/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3380/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..0622b10c6ad050be635175c660db10dd9aec0775
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3380/trainer_state.json
@@ -0,0 +1,3583 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.14475271411339,
+ "eval_steps": 20,
+ "global_step": 3380,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.709091836777472e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-340/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-340/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-340/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-340/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-340/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-340/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-340/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-340/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-340/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-340/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-340/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-340/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-340/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-340/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..3d25234f314befbaf4be05ee5ffa927ba71c3a5d
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-340/trainer_state.json
@@ -0,0 +1,391 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.8202653799758746,
+ "eval_steps": 20,
+ "global_step": 340,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.74929212114944e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3400/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3400/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3400/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3400/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3400/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3400/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3400/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3400/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3400/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3400/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3400/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3400/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3400/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3400/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..9b72e0a8304aeaf88348fc57c5d7c5bf40ff55ae
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3400/trainer_state.json
@@ -0,0 +1,3604 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.193003618817853,
+ "eval_steps": 20,
+ "global_step": 3400,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.72045844350464e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3420/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3420/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3420/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3420/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3420/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3420/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3420/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3420/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3420/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3420/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3420/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3420/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3420/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3420/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..fdf3a36bf3092a1d49e0c48141a290ccc6c70194
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3420/trainer_state.json
@@ -0,0 +1,3625 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.241254523522317,
+ "eval_steps": 20,
+ "global_step": 3420,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.730541425851392e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3440/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3440/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3440/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3440/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3440/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3440/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3440/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3440/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3440/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3440/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3440/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3440/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3440/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3440/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..340a71ba47d21b73ccef431629c99682fc6b880f
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3440/trainer_state.json
@@ -0,0 +1,3646 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.289505428226779,
+ "eval_steps": 20,
+ "global_step": 3440,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.739404383747072e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3460/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3460/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3460/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3460/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3460/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3460/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3460/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3460/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3460/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3460/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3460/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3460/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3460/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3460/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..60ffa21f55145ee7e81cf3469580ec8bfaeae7dd
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3460/trainer_state.json
@@ -0,0 +1,3667 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.337756332931242,
+ "eval_steps": 20,
+ "global_step": 3460,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.74928357277184e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3480/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3480/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3480/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3480/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3480/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3480/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3480/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3480/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3480/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3480/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3480/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3480/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3480/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3480/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..2ca735b863c8d79ebdf7ab58d6dd54245617ec75
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3480/trainer_state.json
@@ -0,0 +1,3688 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.386007237635706,
+ "eval_steps": 20,
+ "global_step": 3480,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.760041535013888e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3500/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3500/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3500/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3500/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3500/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3500/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3500/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3500/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3500/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3500/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3500/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3500/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3500/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3500/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..83de49e937a24b21f0ad18211e18e8cda3f899e2
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3500/trainer_state.json
@@ -0,0 +1,3709 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.43425814234017,
+ "eval_steps": 20,
+ "global_step": 3500,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.770493123402752e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3520/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3520/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3520/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3520/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3520/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3520/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3520/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3520/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3520/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3520/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3520/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3520/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3520/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3520/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..2a5810fd6d1ef219cc44917f5b3dedaf7e250e7a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3520/trainer_state.json
@@ -0,0 +1,3730 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.482509047044632,
+ "eval_steps": 20,
+ "global_step": 3520,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.78057063478784e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3540/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3540/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3540/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3540/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3540/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3540/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3540/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3540/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3540/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3540/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3540/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3540/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3540/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3540/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..9b25c94bec6b726310ef8fed8af9c99d3e654860
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3540/trainer_state.json
@@ -0,0 +1,3751 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.530759951749095,
+ "eval_steps": 20,
+ "global_step": 3540,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.79102701026816e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3560/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3560/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3560/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3560/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3560/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3560/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3560/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3560/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3560/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3560/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3560/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3560/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3560/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3560/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..3637d2e1921c0fda718238a45a8b10952613fa26
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3560/trainer_state.json
@@ -0,0 +1,3772 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.579010856453559,
+ "eval_steps": 20,
+ "global_step": 3560,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.8009636443904e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3580/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3580/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3580/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3580/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3580/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3580/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3580/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3580/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3580/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3580/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3580/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3580/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3580/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3580/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..25a7b852783ddec6b978299c63242c1de674a3b1
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3580/trainer_state.json
@@ -0,0 +1,3793 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.627261761158021,
+ "eval_steps": 20,
+ "global_step": 3580,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.811924716084224e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-360/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-360/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-360/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-360/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-360/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-360/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-360/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-360/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-360/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-360/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-360/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-360/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-360/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-360/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..42694e36d2b200e8deb7f8d093c9ad671056534f
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-360/trainer_state.json
@@ -0,0 +1,412 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.8685162846803377,
+ "eval_steps": 20,
+ "global_step": 360,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.84696930295808e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3600/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3600/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3600/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3600/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3600/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3600/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3600/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3600/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3600/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3600/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3600/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3600/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3600/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3600/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ea7b41319b0a75d484401912df26df1d4beb131
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3600/trainer_state.json
@@ -0,0 +1,3814 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.675512665862485,
+ "eval_steps": 20,
+ "global_step": 3600,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.822174562761728e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3620/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3620/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3620/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3620/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3620/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3620/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3620/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3620/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3620/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3620/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3620/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3620/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3620/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3620/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..7386f84c449841104ebc4daab6296ff216734a9c
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3620/trainer_state.json
@@ -0,0 +1,3835 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.723763570566948,
+ "eval_steps": 20,
+ "global_step": 3620,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.83264051242496e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3640/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3640/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3640/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3640/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3640/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3640/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3640/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3640/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3640/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3640/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3640/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3640/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3640/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3640/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..651a3ff84184562102bde7ef5e7f972227811987
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3640/trainer_state.json
@@ -0,0 +1,3856 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.772014475271412,
+ "eval_steps": 20,
+ "global_step": 3640,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.843316410242048e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3660/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3660/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3660/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3660/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3660/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3660/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3660/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3660/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3660/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3660/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3660/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3660/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3660/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3660/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..a7e67dab46cf4e92f4eb57e9ef15dbc3970cd1f1
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3660/trainer_state.json
@@ -0,0 +1,3877 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.820265379975874,
+ "eval_steps": 20,
+ "global_step": 3660,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.852651922451456e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3680/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3680/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3680/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3680/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3680/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3680/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3680/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3680/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3680/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3680/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3680/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3680/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3680/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3680/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..3e1fb13da59a0bed52c065758673d281fab6ee4f
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3680/trainer_state.json
@@ -0,0 +1,3898 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.868516284680338,
+ "eval_steps": 20,
+ "global_step": 3680,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ },
+ {
+ "entropy": 0.0912214593961835,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.16693317890167236,
+ "learning_rate": 1.6744016157207948e-05,
+ "loss": 0.06410449743270874,
+ "mean_token_accuracy": 0.9774021029472351,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.18701685898089676,
+ "eval_loss": 1.1080079078674316,
+ "eval_mean_token_accuracy": 0.8379885400279184,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 100.8862,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3680
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.863077523772416e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3700/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3700/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3700/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3700/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3700/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3700/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3700/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3700/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3700/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3700/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3700/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3700/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3700/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3700/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..0c182092fb8bec7f7cbd01efe4abc416c4d1c364
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3700/trainer_state.json
@@ -0,0 +1,3919 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.916767189384801,
+ "eval_steps": 20,
+ "global_step": 3700,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ },
+ {
+ "entropy": 0.0912214593961835,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.16693317890167236,
+ "learning_rate": 1.6744016157207948e-05,
+ "loss": 0.06410449743270874,
+ "mean_token_accuracy": 0.9774021029472351,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.18701685898089676,
+ "eval_loss": 1.1080079078674316,
+ "eval_mean_token_accuracy": 0.8379885400279184,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 100.8862,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3680
+ },
+ {
+ "entropy": 0.09622923573479056,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.15017707645893097,
+ "learning_rate": 1.53689516610855e-05,
+ "loss": 0.06397625803947449,
+ "mean_token_accuracy": 0.9777478039264679,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.1863192869873529,
+ "eval_loss": 1.1090794801712036,
+ "eval_mean_token_accuracy": 0.8382032892007506,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 100.8767,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3700
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.872863022578688e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3720/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3720/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3720/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3720/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3720/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3720/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3720/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3720/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3720/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3720/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3720/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3720/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3720/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3720/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..7713a819887930c6e68732705d8edde52a923eb0
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3720/trainer_state.json
@@ -0,0 +1,3940 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 8.965018094089265,
+ "eval_steps": 20,
+ "global_step": 3720,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ },
+ {
+ "entropy": 0.0912214593961835,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.16693317890167236,
+ "learning_rate": 1.6744016157207948e-05,
+ "loss": 0.06410449743270874,
+ "mean_token_accuracy": 0.9774021029472351,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.18701685898089676,
+ "eval_loss": 1.1080079078674316,
+ "eval_mean_token_accuracy": 0.8379885400279184,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 100.8862,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3680
+ },
+ {
+ "entropy": 0.09622923573479056,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.15017707645893097,
+ "learning_rate": 1.53689516610855e-05,
+ "loss": 0.06397625803947449,
+ "mean_token_accuracy": 0.9777478039264679,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.1863192869873529,
+ "eval_loss": 1.1090794801712036,
+ "eval_mean_token_accuracy": 0.8382032892007506,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 100.8767,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3700
+ },
+ {
+ "entropy": 0.0912615241482854,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.1881953328847885,
+ "learning_rate": 1.4050714515625077e-05,
+ "loss": 0.062059921026229856,
+ "mean_token_accuracy": 0.9778110668063164,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.18606430442815416,
+ "eval_loss": 1.109899640083313,
+ "eval_mean_token_accuracy": 0.8379705333977603,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 100.7437,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3720
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.883224340100096e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3740/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3740/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3740/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3740/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3740/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3740/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3740/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3740/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3740/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3740/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3740/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3740/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3740/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3740/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..0c80a95ab9da1d4e4600a20606ab97f9d2fa5fe2
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3740/trainer_state.json
@@ -0,0 +1,3961 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 9.012062726176115,
+ "eval_steps": 20,
+ "global_step": 3740,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ },
+ {
+ "entropy": 0.0912214593961835,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.16693317890167236,
+ "learning_rate": 1.6744016157207948e-05,
+ "loss": 0.06410449743270874,
+ "mean_token_accuracy": 0.9774021029472351,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.18701685898089676,
+ "eval_loss": 1.1080079078674316,
+ "eval_mean_token_accuracy": 0.8379885400279184,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 100.8862,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3680
+ },
+ {
+ "entropy": 0.09622923573479056,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.15017707645893097,
+ "learning_rate": 1.53689516610855e-05,
+ "loss": 0.06397625803947449,
+ "mean_token_accuracy": 0.9777478039264679,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.1863192869873529,
+ "eval_loss": 1.1090794801712036,
+ "eval_mean_token_accuracy": 0.8382032892007506,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 100.8767,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3700
+ },
+ {
+ "entropy": 0.0912615241482854,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.1881953328847885,
+ "learning_rate": 1.4050714515625077e-05,
+ "loss": 0.062059921026229856,
+ "mean_token_accuracy": 0.9778110668063164,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.18606430442815416,
+ "eval_loss": 1.109899640083313,
+ "eval_mean_token_accuracy": 0.8379705333977603,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 100.7437,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3720
+ },
+ {
+ "entropy": 0.09705390800268222,
+ "epoch": 9.012062726176115,
+ "grad_norm": 0.3272952437400818,
+ "learning_rate": 1.2789677766345677e-05,
+ "loss": 0.06879411935806275,
+ "mean_token_accuracy": 0.9774823463880099,
+ "num_tokens": 9122597.0,
+ "step": 3740
+ },
+ {
+ "epoch": 9.012062726176115,
+ "eval_entropy": 0.18716304565078756,
+ "eval_loss": 1.103650689125061,
+ "eval_mean_token_accuracy": 0.8379069282097763,
+ "eval_num_tokens": 9122597.0,
+ "eval_runtime": 100.8591,
+ "eval_samples_per_second": 14.079,
+ "eval_steps_per_second": 1.765,
+ "step": 3740
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.892632342551552e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3760/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3760/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3760/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3760/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3760/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3760/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3760/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3760/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3760/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3760/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3760/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3760/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3760/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3760/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..206cd8a33837fe717b27d9bc6e7c8ef1b5e7010e
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3760/trainer_state.json
@@ -0,0 +1,3982 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 9.060313630880579,
+ "eval_steps": 20,
+ "global_step": 3760,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ },
+ {
+ "entropy": 0.0912214593961835,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.16693317890167236,
+ "learning_rate": 1.6744016157207948e-05,
+ "loss": 0.06410449743270874,
+ "mean_token_accuracy": 0.9774021029472351,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.18701685898089676,
+ "eval_loss": 1.1080079078674316,
+ "eval_mean_token_accuracy": 0.8379885400279184,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 100.8862,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3680
+ },
+ {
+ "entropy": 0.09622923573479056,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.15017707645893097,
+ "learning_rate": 1.53689516610855e-05,
+ "loss": 0.06397625803947449,
+ "mean_token_accuracy": 0.9777478039264679,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.1863192869873529,
+ "eval_loss": 1.1090794801712036,
+ "eval_mean_token_accuracy": 0.8382032892007506,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 100.8767,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3700
+ },
+ {
+ "entropy": 0.0912615241482854,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.1881953328847885,
+ "learning_rate": 1.4050714515625077e-05,
+ "loss": 0.062059921026229856,
+ "mean_token_accuracy": 0.9778110668063164,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.18606430442815416,
+ "eval_loss": 1.109899640083313,
+ "eval_mean_token_accuracy": 0.8379705333977603,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 100.7437,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3720
+ },
+ {
+ "entropy": 0.09705390800268222,
+ "epoch": 9.012062726176115,
+ "grad_norm": 0.3272952437400818,
+ "learning_rate": 1.2789677766345677e-05,
+ "loss": 0.06879411935806275,
+ "mean_token_accuracy": 0.9774823463880099,
+ "num_tokens": 9122597.0,
+ "step": 3740
+ },
+ {
+ "epoch": 9.012062726176115,
+ "eval_entropy": 0.18716304565078756,
+ "eval_loss": 1.103650689125061,
+ "eval_mean_token_accuracy": 0.8379069282097763,
+ "eval_num_tokens": 9122597.0,
+ "eval_runtime": 100.8591,
+ "eval_samples_per_second": 14.079,
+ "eval_steps_per_second": 1.765,
+ "step": 3740
+ },
+ {
+ "entropy": 0.09239197438582777,
+ "epoch": 9.060313630880579,
+ "grad_norm": 0.12848101556301117,
+ "learning_rate": 1.1586198271730714e-05,
+ "loss": 0.058521777391433716,
+ "mean_token_accuracy": 0.9808280184864998,
+ "num_tokens": 9171477.0,
+ "step": 3760
+ },
+ {
+ "epoch": 9.060313630880579,
+ "eval_entropy": 0.18443025656965342,
+ "eval_loss": 1.1128716468811035,
+ "eval_mean_token_accuracy": 0.8376701189560837,
+ "eval_num_tokens": 9171477.0,
+ "eval_runtime": 100.7759,
+ "eval_samples_per_second": 14.091,
+ "eval_steps_per_second": 1.766,
+ "step": 3760
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.902099841711104e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3780/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3780/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3780/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3780/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3780/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3780/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3780/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3780/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3780/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3780/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3780/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3780/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3780/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3780/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..d78a8f5c5f80f81d50c39c79a4a0247aa5eec023
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3780/trainer_state.json
@@ -0,0 +1,4003 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 9.108564535585042,
+ "eval_steps": 20,
+ "global_step": 3780,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ },
+ {
+ "entropy": 0.0912214593961835,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.16693317890167236,
+ "learning_rate": 1.6744016157207948e-05,
+ "loss": 0.06410449743270874,
+ "mean_token_accuracy": 0.9774021029472351,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.18701685898089676,
+ "eval_loss": 1.1080079078674316,
+ "eval_mean_token_accuracy": 0.8379885400279184,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 100.8862,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3680
+ },
+ {
+ "entropy": 0.09622923573479056,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.15017707645893097,
+ "learning_rate": 1.53689516610855e-05,
+ "loss": 0.06397625803947449,
+ "mean_token_accuracy": 0.9777478039264679,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.1863192869873529,
+ "eval_loss": 1.1090794801712036,
+ "eval_mean_token_accuracy": 0.8382032892007506,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 100.8767,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3700
+ },
+ {
+ "entropy": 0.0912615241482854,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.1881953328847885,
+ "learning_rate": 1.4050714515625077e-05,
+ "loss": 0.062059921026229856,
+ "mean_token_accuracy": 0.9778110668063164,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.18606430442815416,
+ "eval_loss": 1.109899640083313,
+ "eval_mean_token_accuracy": 0.8379705333977603,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 100.7437,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3720
+ },
+ {
+ "entropy": 0.09705390800268222,
+ "epoch": 9.012062726176115,
+ "grad_norm": 0.3272952437400818,
+ "learning_rate": 1.2789677766345677e-05,
+ "loss": 0.06879411935806275,
+ "mean_token_accuracy": 0.9774823463880099,
+ "num_tokens": 9122597.0,
+ "step": 3740
+ },
+ {
+ "epoch": 9.012062726176115,
+ "eval_entropy": 0.18716304565078756,
+ "eval_loss": 1.103650689125061,
+ "eval_mean_token_accuracy": 0.8379069282097763,
+ "eval_num_tokens": 9122597.0,
+ "eval_runtime": 100.8591,
+ "eval_samples_per_second": 14.079,
+ "eval_steps_per_second": 1.765,
+ "step": 3740
+ },
+ {
+ "entropy": 0.09239197438582777,
+ "epoch": 9.060313630880579,
+ "grad_norm": 0.12848101556301117,
+ "learning_rate": 1.1586198271730714e-05,
+ "loss": 0.058521777391433716,
+ "mean_token_accuracy": 0.9808280184864998,
+ "num_tokens": 9171477.0,
+ "step": 3760
+ },
+ {
+ "epoch": 9.060313630880579,
+ "eval_entropy": 0.18443025656965342,
+ "eval_loss": 1.1128716468811035,
+ "eval_mean_token_accuracy": 0.8376701189560837,
+ "eval_num_tokens": 9171477.0,
+ "eval_runtime": 100.7759,
+ "eval_samples_per_second": 14.091,
+ "eval_steps_per_second": 1.766,
+ "step": 3760
+ },
+ {
+ "entropy": 0.09276993470266462,
+ "epoch": 9.108564535585042,
+ "grad_norm": 0.0987873300909996,
+ "learning_rate": 1.0440616602241303e-05,
+ "loss": 0.06252194046974183,
+ "mean_token_accuracy": 0.9791185334324837,
+ "num_tokens": 9217543.0,
+ "step": 3780
+ },
+ {
+ "epoch": 9.108564535585042,
+ "eval_entropy": 0.18163840397355263,
+ "eval_loss": 1.124428629875183,
+ "eval_mean_token_accuracy": 0.8379190537366974,
+ "eval_num_tokens": 9217543.0,
+ "eval_runtime": 100.7229,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3780
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.911165225188352e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-380/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-380/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-380/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-380/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-380/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-380/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-380/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-380/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-380/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-380/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-380/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-380/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-380/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-380/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..5e5f4133641e526a605a58321cc34853d3d8c6a9
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-380/trainer_state.json
@@ -0,0 +1,433 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.916767189384801,
+ "eval_steps": 20,
+ "global_step": 380,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.93760262162432e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3800/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3800/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3800/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3800/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3800/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3800/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3800/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3800/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3800/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3800/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3800/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3800/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3800/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3800/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..feba21023919f362a241e7ec76db9c2803a1b0bf
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3800/trainer_state.json
@@ -0,0 +1,4024 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 9.156815440289506,
+ "eval_steps": 20,
+ "global_step": 3800,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ },
+ {
+ "entropy": 0.0912214593961835,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.16693317890167236,
+ "learning_rate": 1.6744016157207948e-05,
+ "loss": 0.06410449743270874,
+ "mean_token_accuracy": 0.9774021029472351,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.18701685898089676,
+ "eval_loss": 1.1080079078674316,
+ "eval_mean_token_accuracy": 0.8379885400279184,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 100.8862,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3680
+ },
+ {
+ "entropy": 0.09622923573479056,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.15017707645893097,
+ "learning_rate": 1.53689516610855e-05,
+ "loss": 0.06397625803947449,
+ "mean_token_accuracy": 0.9777478039264679,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.1863192869873529,
+ "eval_loss": 1.1090794801712036,
+ "eval_mean_token_accuracy": 0.8382032892007506,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 100.8767,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3700
+ },
+ {
+ "entropy": 0.0912615241482854,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.1881953328847885,
+ "learning_rate": 1.4050714515625077e-05,
+ "loss": 0.062059921026229856,
+ "mean_token_accuracy": 0.9778110668063164,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.18606430442815416,
+ "eval_loss": 1.109899640083313,
+ "eval_mean_token_accuracy": 0.8379705333977603,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 100.7437,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3720
+ },
+ {
+ "entropy": 0.09705390800268222,
+ "epoch": 9.012062726176115,
+ "grad_norm": 0.3272952437400818,
+ "learning_rate": 1.2789677766345677e-05,
+ "loss": 0.06879411935806275,
+ "mean_token_accuracy": 0.9774823463880099,
+ "num_tokens": 9122597.0,
+ "step": 3740
+ },
+ {
+ "epoch": 9.012062726176115,
+ "eval_entropy": 0.18716304565078756,
+ "eval_loss": 1.103650689125061,
+ "eval_mean_token_accuracy": 0.8379069282097763,
+ "eval_num_tokens": 9122597.0,
+ "eval_runtime": 100.8591,
+ "eval_samples_per_second": 14.079,
+ "eval_steps_per_second": 1.765,
+ "step": 3740
+ },
+ {
+ "entropy": 0.09239197438582777,
+ "epoch": 9.060313630880579,
+ "grad_norm": 0.12848101556301117,
+ "learning_rate": 1.1586198271730714e-05,
+ "loss": 0.058521777391433716,
+ "mean_token_accuracy": 0.9808280184864998,
+ "num_tokens": 9171477.0,
+ "step": 3760
+ },
+ {
+ "epoch": 9.060313630880579,
+ "eval_entropy": 0.18443025656965342,
+ "eval_loss": 1.1128716468811035,
+ "eval_mean_token_accuracy": 0.8376701189560837,
+ "eval_num_tokens": 9171477.0,
+ "eval_runtime": 100.7759,
+ "eval_samples_per_second": 14.091,
+ "eval_steps_per_second": 1.766,
+ "step": 3760
+ },
+ {
+ "entropy": 0.09276993470266462,
+ "epoch": 9.108564535585042,
+ "grad_norm": 0.0987873300909996,
+ "learning_rate": 1.0440616602241303e-05,
+ "loss": 0.06252194046974183,
+ "mean_token_accuracy": 0.9791185334324837,
+ "num_tokens": 9217543.0,
+ "step": 3780
+ },
+ {
+ "epoch": 9.108564535585042,
+ "eval_entropy": 0.18163840397355263,
+ "eval_loss": 1.124428629875183,
+ "eval_mean_token_accuracy": 0.8379190537366974,
+ "eval_num_tokens": 9217543.0,
+ "eval_runtime": 100.7229,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3780
+ },
+ {
+ "entropy": 0.08816529829055071,
+ "epoch": 9.156815440289506,
+ "grad_norm": 0.10476374626159668,
+ "learning_rate": 9.353256943939065e-06,
+ "loss": 0.056501579284667966,
+ "mean_token_accuracy": 0.9800700470805168,
+ "num_tokens": 9268724.0,
+ "step": 3800
+ },
+ {
+ "epoch": 9.156815440289506,
+ "eval_entropy": 0.17974376218037658,
+ "eval_loss": 1.1325938701629639,
+ "eval_mean_token_accuracy": 0.8377991332097,
+ "eval_num_tokens": 9268724.0,
+ "eval_runtime": 100.8729,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3800
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.922020296999936e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3820/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3820/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3820/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3820/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3820/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3820/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3820/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3820/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3820/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3820/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3820/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3820/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3820/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3820/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..f51589026a3e1a688cb415b4e8efc3ed966f7318
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3820/trainer_state.json
@@ -0,0 +1,4045 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 9.205066344993968,
+ "eval_steps": 20,
+ "global_step": 3820,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ },
+ {
+ "entropy": 0.0912214593961835,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.16693317890167236,
+ "learning_rate": 1.6744016157207948e-05,
+ "loss": 0.06410449743270874,
+ "mean_token_accuracy": 0.9774021029472351,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.18701685898089676,
+ "eval_loss": 1.1080079078674316,
+ "eval_mean_token_accuracy": 0.8379885400279184,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 100.8862,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3680
+ },
+ {
+ "entropy": 0.09622923573479056,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.15017707645893097,
+ "learning_rate": 1.53689516610855e-05,
+ "loss": 0.06397625803947449,
+ "mean_token_accuracy": 0.9777478039264679,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.1863192869873529,
+ "eval_loss": 1.1090794801712036,
+ "eval_mean_token_accuracy": 0.8382032892007506,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 100.8767,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3700
+ },
+ {
+ "entropy": 0.0912615241482854,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.1881953328847885,
+ "learning_rate": 1.4050714515625077e-05,
+ "loss": 0.062059921026229856,
+ "mean_token_accuracy": 0.9778110668063164,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.18606430442815416,
+ "eval_loss": 1.109899640083313,
+ "eval_mean_token_accuracy": 0.8379705333977603,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 100.7437,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3720
+ },
+ {
+ "entropy": 0.09705390800268222,
+ "epoch": 9.012062726176115,
+ "grad_norm": 0.3272952437400818,
+ "learning_rate": 1.2789677766345677e-05,
+ "loss": 0.06879411935806275,
+ "mean_token_accuracy": 0.9774823463880099,
+ "num_tokens": 9122597.0,
+ "step": 3740
+ },
+ {
+ "epoch": 9.012062726176115,
+ "eval_entropy": 0.18716304565078756,
+ "eval_loss": 1.103650689125061,
+ "eval_mean_token_accuracy": 0.8379069282097763,
+ "eval_num_tokens": 9122597.0,
+ "eval_runtime": 100.8591,
+ "eval_samples_per_second": 14.079,
+ "eval_steps_per_second": 1.765,
+ "step": 3740
+ },
+ {
+ "entropy": 0.09239197438582777,
+ "epoch": 9.060313630880579,
+ "grad_norm": 0.12848101556301117,
+ "learning_rate": 1.1586198271730714e-05,
+ "loss": 0.058521777391433716,
+ "mean_token_accuracy": 0.9808280184864998,
+ "num_tokens": 9171477.0,
+ "step": 3760
+ },
+ {
+ "epoch": 9.060313630880579,
+ "eval_entropy": 0.18443025656965342,
+ "eval_loss": 1.1128716468811035,
+ "eval_mean_token_accuracy": 0.8376701189560837,
+ "eval_num_tokens": 9171477.0,
+ "eval_runtime": 100.7759,
+ "eval_samples_per_second": 14.091,
+ "eval_steps_per_second": 1.766,
+ "step": 3760
+ },
+ {
+ "entropy": 0.09276993470266462,
+ "epoch": 9.108564535585042,
+ "grad_norm": 0.0987873300909996,
+ "learning_rate": 1.0440616602241303e-05,
+ "loss": 0.06252194046974183,
+ "mean_token_accuracy": 0.9791185334324837,
+ "num_tokens": 9217543.0,
+ "step": 3780
+ },
+ {
+ "epoch": 9.108564535585042,
+ "eval_entropy": 0.18163840397355263,
+ "eval_loss": 1.124428629875183,
+ "eval_mean_token_accuracy": 0.8379190537366974,
+ "eval_num_tokens": 9217543.0,
+ "eval_runtime": 100.7229,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3780
+ },
+ {
+ "entropy": 0.08816529829055071,
+ "epoch": 9.156815440289506,
+ "grad_norm": 0.10476374626159668,
+ "learning_rate": 9.353256943939065e-06,
+ "loss": 0.056501579284667966,
+ "mean_token_accuracy": 0.9800700470805168,
+ "num_tokens": 9268724.0,
+ "step": 3800
+ },
+ {
+ "epoch": 9.156815440289506,
+ "eval_entropy": 0.17974376218037658,
+ "eval_loss": 1.1325938701629639,
+ "eval_mean_token_accuracy": 0.8377991332097,
+ "eval_num_tokens": 9268724.0,
+ "eval_runtime": 100.8729,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3800
+ },
+ {
+ "entropy": 0.0886100891046226,
+ "epoch": 9.205066344993968,
+ "grad_norm": 0.10369760543107986,
+ "learning_rate": 8.32442700674502e-06,
+ "loss": 0.05948063135147095,
+ "mean_token_accuracy": 0.9798692613840103,
+ "num_tokens": 9316805.0,
+ "step": 3820
+ },
+ {
+ "epoch": 9.205066344993968,
+ "eval_entropy": 0.1793072780364015,
+ "eval_loss": 1.1363310813903809,
+ "eval_mean_token_accuracy": 0.8378076965219519,
+ "eval_num_tokens": 9316805.0,
+ "eval_runtime": 100.8909,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3820
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.932807665660928e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3840/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3840/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3840/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3840/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3840/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3840/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3840/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3840/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3840/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3840/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3840/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3840/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3840/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3840/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..a74da7aaaae4ee351f58ee3164f52e2ac28c8a59
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3840/trainer_state.json
@@ -0,0 +1,4066 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 9.253317249698432,
+ "eval_steps": 20,
+ "global_step": 3840,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ },
+ {
+ "entropy": 0.0912214593961835,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.16693317890167236,
+ "learning_rate": 1.6744016157207948e-05,
+ "loss": 0.06410449743270874,
+ "mean_token_accuracy": 0.9774021029472351,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.18701685898089676,
+ "eval_loss": 1.1080079078674316,
+ "eval_mean_token_accuracy": 0.8379885400279184,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 100.8862,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3680
+ },
+ {
+ "entropy": 0.09622923573479056,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.15017707645893097,
+ "learning_rate": 1.53689516610855e-05,
+ "loss": 0.06397625803947449,
+ "mean_token_accuracy": 0.9777478039264679,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.1863192869873529,
+ "eval_loss": 1.1090794801712036,
+ "eval_mean_token_accuracy": 0.8382032892007506,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 100.8767,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3700
+ },
+ {
+ "entropy": 0.0912615241482854,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.1881953328847885,
+ "learning_rate": 1.4050714515625077e-05,
+ "loss": 0.062059921026229856,
+ "mean_token_accuracy": 0.9778110668063164,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.18606430442815416,
+ "eval_loss": 1.109899640083313,
+ "eval_mean_token_accuracy": 0.8379705333977603,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 100.7437,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3720
+ },
+ {
+ "entropy": 0.09705390800268222,
+ "epoch": 9.012062726176115,
+ "grad_norm": 0.3272952437400818,
+ "learning_rate": 1.2789677766345677e-05,
+ "loss": 0.06879411935806275,
+ "mean_token_accuracy": 0.9774823463880099,
+ "num_tokens": 9122597.0,
+ "step": 3740
+ },
+ {
+ "epoch": 9.012062726176115,
+ "eval_entropy": 0.18716304565078756,
+ "eval_loss": 1.103650689125061,
+ "eval_mean_token_accuracy": 0.8379069282097763,
+ "eval_num_tokens": 9122597.0,
+ "eval_runtime": 100.8591,
+ "eval_samples_per_second": 14.079,
+ "eval_steps_per_second": 1.765,
+ "step": 3740
+ },
+ {
+ "entropy": 0.09239197438582777,
+ "epoch": 9.060313630880579,
+ "grad_norm": 0.12848101556301117,
+ "learning_rate": 1.1586198271730714e-05,
+ "loss": 0.058521777391433716,
+ "mean_token_accuracy": 0.9808280184864998,
+ "num_tokens": 9171477.0,
+ "step": 3760
+ },
+ {
+ "epoch": 9.060313630880579,
+ "eval_entropy": 0.18443025656965342,
+ "eval_loss": 1.1128716468811035,
+ "eval_mean_token_accuracy": 0.8376701189560837,
+ "eval_num_tokens": 9171477.0,
+ "eval_runtime": 100.7759,
+ "eval_samples_per_second": 14.091,
+ "eval_steps_per_second": 1.766,
+ "step": 3760
+ },
+ {
+ "entropy": 0.09276993470266462,
+ "epoch": 9.108564535585042,
+ "grad_norm": 0.0987873300909996,
+ "learning_rate": 1.0440616602241303e-05,
+ "loss": 0.06252194046974183,
+ "mean_token_accuracy": 0.9791185334324837,
+ "num_tokens": 9217543.0,
+ "step": 3780
+ },
+ {
+ "epoch": 9.108564535585042,
+ "eval_entropy": 0.18163840397355263,
+ "eval_loss": 1.124428629875183,
+ "eval_mean_token_accuracy": 0.8379190537366974,
+ "eval_num_tokens": 9217543.0,
+ "eval_runtime": 100.7229,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3780
+ },
+ {
+ "entropy": 0.08816529829055071,
+ "epoch": 9.156815440289506,
+ "grad_norm": 0.10476374626159668,
+ "learning_rate": 9.353256943939065e-06,
+ "loss": 0.056501579284667966,
+ "mean_token_accuracy": 0.9800700470805168,
+ "num_tokens": 9268724.0,
+ "step": 3800
+ },
+ {
+ "epoch": 9.156815440289506,
+ "eval_entropy": 0.17974376218037658,
+ "eval_loss": 1.1325938701629639,
+ "eval_mean_token_accuracy": 0.8377991332097,
+ "eval_num_tokens": 9268724.0,
+ "eval_runtime": 100.8729,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3800
+ },
+ {
+ "entropy": 0.0886100891046226,
+ "epoch": 9.205066344993968,
+ "grad_norm": 0.10369760543107986,
+ "learning_rate": 8.32442700674502e-06,
+ "loss": 0.05948063135147095,
+ "mean_token_accuracy": 0.9798692613840103,
+ "num_tokens": 9316805.0,
+ "step": 3820
+ },
+ {
+ "epoch": 9.205066344993968,
+ "eval_entropy": 0.1793072780364015,
+ "eval_loss": 1.1363310813903809,
+ "eval_mean_token_accuracy": 0.8378076965219519,
+ "eval_num_tokens": 9316805.0,
+ "eval_runtime": 100.8909,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3820
+ },
+ {
+ "entropy": 0.07910045385360717,
+ "epoch": 9.253317249698432,
+ "grad_norm": 0.06480544060468674,
+ "learning_rate": 7.3544179373615505e-06,
+ "loss": 0.053755837678909305,
+ "mean_token_accuracy": 0.9819905251264572,
+ "num_tokens": 9369940.0,
+ "step": 3840
+ },
+ {
+ "epoch": 9.253317249698432,
+ "eval_entropy": 0.1782708638086078,
+ "eval_loss": 1.1404398679733276,
+ "eval_mean_token_accuracy": 0.8379109310969878,
+ "eval_num_tokens": 9369940.0,
+ "eval_runtime": 100.7264,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3840
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.943333795902464e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3860/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3860/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3860/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3860/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3860/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3860/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3860/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3860/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3860/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3860/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3860/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3860/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3860/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3860/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..cec593ff929f55b8d910f559eb6907c3a0652578
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3860/trainer_state.json
@@ -0,0 +1,4087 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 9.301568154402895,
+ "eval_steps": 20,
+ "global_step": 3860,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ },
+ {
+ "entropy": 0.0912214593961835,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.16693317890167236,
+ "learning_rate": 1.6744016157207948e-05,
+ "loss": 0.06410449743270874,
+ "mean_token_accuracy": 0.9774021029472351,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.18701685898089676,
+ "eval_loss": 1.1080079078674316,
+ "eval_mean_token_accuracy": 0.8379885400279184,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 100.8862,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3680
+ },
+ {
+ "entropy": 0.09622923573479056,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.15017707645893097,
+ "learning_rate": 1.53689516610855e-05,
+ "loss": 0.06397625803947449,
+ "mean_token_accuracy": 0.9777478039264679,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.1863192869873529,
+ "eval_loss": 1.1090794801712036,
+ "eval_mean_token_accuracy": 0.8382032892007506,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 100.8767,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3700
+ },
+ {
+ "entropy": 0.0912615241482854,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.1881953328847885,
+ "learning_rate": 1.4050714515625077e-05,
+ "loss": 0.062059921026229856,
+ "mean_token_accuracy": 0.9778110668063164,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.18606430442815416,
+ "eval_loss": 1.109899640083313,
+ "eval_mean_token_accuracy": 0.8379705333977603,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 100.7437,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3720
+ },
+ {
+ "entropy": 0.09705390800268222,
+ "epoch": 9.012062726176115,
+ "grad_norm": 0.3272952437400818,
+ "learning_rate": 1.2789677766345677e-05,
+ "loss": 0.06879411935806275,
+ "mean_token_accuracy": 0.9774823463880099,
+ "num_tokens": 9122597.0,
+ "step": 3740
+ },
+ {
+ "epoch": 9.012062726176115,
+ "eval_entropy": 0.18716304565078756,
+ "eval_loss": 1.103650689125061,
+ "eval_mean_token_accuracy": 0.8379069282097763,
+ "eval_num_tokens": 9122597.0,
+ "eval_runtime": 100.8591,
+ "eval_samples_per_second": 14.079,
+ "eval_steps_per_second": 1.765,
+ "step": 3740
+ },
+ {
+ "entropy": 0.09239197438582777,
+ "epoch": 9.060313630880579,
+ "grad_norm": 0.12848101556301117,
+ "learning_rate": 1.1586198271730714e-05,
+ "loss": 0.058521777391433716,
+ "mean_token_accuracy": 0.9808280184864998,
+ "num_tokens": 9171477.0,
+ "step": 3760
+ },
+ {
+ "epoch": 9.060313630880579,
+ "eval_entropy": 0.18443025656965342,
+ "eval_loss": 1.1128716468811035,
+ "eval_mean_token_accuracy": 0.8376701189560837,
+ "eval_num_tokens": 9171477.0,
+ "eval_runtime": 100.7759,
+ "eval_samples_per_second": 14.091,
+ "eval_steps_per_second": 1.766,
+ "step": 3760
+ },
+ {
+ "entropy": 0.09276993470266462,
+ "epoch": 9.108564535585042,
+ "grad_norm": 0.0987873300909996,
+ "learning_rate": 1.0440616602241303e-05,
+ "loss": 0.06252194046974183,
+ "mean_token_accuracy": 0.9791185334324837,
+ "num_tokens": 9217543.0,
+ "step": 3780
+ },
+ {
+ "epoch": 9.108564535585042,
+ "eval_entropy": 0.18163840397355263,
+ "eval_loss": 1.124428629875183,
+ "eval_mean_token_accuracy": 0.8379190537366974,
+ "eval_num_tokens": 9217543.0,
+ "eval_runtime": 100.7229,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3780
+ },
+ {
+ "entropy": 0.08816529829055071,
+ "epoch": 9.156815440289506,
+ "grad_norm": 0.10476374626159668,
+ "learning_rate": 9.353256943939065e-06,
+ "loss": 0.056501579284667966,
+ "mean_token_accuracy": 0.9800700470805168,
+ "num_tokens": 9268724.0,
+ "step": 3800
+ },
+ {
+ "epoch": 9.156815440289506,
+ "eval_entropy": 0.17974376218037658,
+ "eval_loss": 1.1325938701629639,
+ "eval_mean_token_accuracy": 0.8377991332097,
+ "eval_num_tokens": 9268724.0,
+ "eval_runtime": 100.8729,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3800
+ },
+ {
+ "entropy": 0.0886100891046226,
+ "epoch": 9.205066344993968,
+ "grad_norm": 0.10369760543107986,
+ "learning_rate": 8.32442700674502e-06,
+ "loss": 0.05948063135147095,
+ "mean_token_accuracy": 0.9798692613840103,
+ "num_tokens": 9316805.0,
+ "step": 3820
+ },
+ {
+ "epoch": 9.205066344993968,
+ "eval_entropy": 0.1793072780364015,
+ "eval_loss": 1.1363310813903809,
+ "eval_mean_token_accuracy": 0.8378076965219519,
+ "eval_num_tokens": 9316805.0,
+ "eval_runtime": 100.8909,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3820
+ },
+ {
+ "entropy": 0.07910045385360717,
+ "epoch": 9.253317249698432,
+ "grad_norm": 0.06480544060468674,
+ "learning_rate": 7.3544179373615505e-06,
+ "loss": 0.053755837678909305,
+ "mean_token_accuracy": 0.9819905251264572,
+ "num_tokens": 9369940.0,
+ "step": 3840
+ },
+ {
+ "epoch": 9.253317249698432,
+ "eval_entropy": 0.1782708638086078,
+ "eval_loss": 1.1404398679733276,
+ "eval_mean_token_accuracy": 0.8379109310969878,
+ "eval_num_tokens": 9369940.0,
+ "eval_runtime": 100.7264,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3840
+ },
+ {
+ "entropy": 0.09386134278029204,
+ "epoch": 9.301568154402895,
+ "grad_norm": 0.12832233309745789,
+ "learning_rate": 6.443504236881163e-06,
+ "loss": 0.061395853757858276,
+ "mean_token_accuracy": 0.9780228048563003,
+ "num_tokens": 9416788.0,
+ "step": 3860
+ },
+ {
+ "epoch": 9.301568154402895,
+ "eval_entropy": 0.1777799724276816,
+ "eval_loss": 1.1431900262832642,
+ "eval_mean_token_accuracy": 0.8377661607908399,
+ "eval_num_tokens": 9416788.0,
+ "eval_runtime": 100.8637,
+ "eval_samples_per_second": 14.078,
+ "eval_steps_per_second": 1.765,
+ "step": 3860
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.953259488101376e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3880/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3880/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3880/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3880/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3880/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3880/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3880/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3880/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3880/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3880/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3880/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3880/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3880/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3880/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..4910bd5f3194ffb01addd3908ab6e2d8b73e321a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3880/trainer_state.json
@@ -0,0 +1,4108 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 9.349819059107359,
+ "eval_steps": 20,
+ "global_step": 3880,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ },
+ {
+ "entropy": 0.0912214593961835,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.16693317890167236,
+ "learning_rate": 1.6744016157207948e-05,
+ "loss": 0.06410449743270874,
+ "mean_token_accuracy": 0.9774021029472351,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.18701685898089676,
+ "eval_loss": 1.1080079078674316,
+ "eval_mean_token_accuracy": 0.8379885400279184,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 100.8862,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3680
+ },
+ {
+ "entropy": 0.09622923573479056,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.15017707645893097,
+ "learning_rate": 1.53689516610855e-05,
+ "loss": 0.06397625803947449,
+ "mean_token_accuracy": 0.9777478039264679,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.1863192869873529,
+ "eval_loss": 1.1090794801712036,
+ "eval_mean_token_accuracy": 0.8382032892007506,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 100.8767,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3700
+ },
+ {
+ "entropy": 0.0912615241482854,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.1881953328847885,
+ "learning_rate": 1.4050714515625077e-05,
+ "loss": 0.062059921026229856,
+ "mean_token_accuracy": 0.9778110668063164,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.18606430442815416,
+ "eval_loss": 1.109899640083313,
+ "eval_mean_token_accuracy": 0.8379705333977603,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 100.7437,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3720
+ },
+ {
+ "entropy": 0.09705390800268222,
+ "epoch": 9.012062726176115,
+ "grad_norm": 0.3272952437400818,
+ "learning_rate": 1.2789677766345677e-05,
+ "loss": 0.06879411935806275,
+ "mean_token_accuracy": 0.9774823463880099,
+ "num_tokens": 9122597.0,
+ "step": 3740
+ },
+ {
+ "epoch": 9.012062726176115,
+ "eval_entropy": 0.18716304565078756,
+ "eval_loss": 1.103650689125061,
+ "eval_mean_token_accuracy": 0.8379069282097763,
+ "eval_num_tokens": 9122597.0,
+ "eval_runtime": 100.8591,
+ "eval_samples_per_second": 14.079,
+ "eval_steps_per_second": 1.765,
+ "step": 3740
+ },
+ {
+ "entropy": 0.09239197438582777,
+ "epoch": 9.060313630880579,
+ "grad_norm": 0.12848101556301117,
+ "learning_rate": 1.1586198271730714e-05,
+ "loss": 0.058521777391433716,
+ "mean_token_accuracy": 0.9808280184864998,
+ "num_tokens": 9171477.0,
+ "step": 3760
+ },
+ {
+ "epoch": 9.060313630880579,
+ "eval_entropy": 0.18443025656965342,
+ "eval_loss": 1.1128716468811035,
+ "eval_mean_token_accuracy": 0.8376701189560837,
+ "eval_num_tokens": 9171477.0,
+ "eval_runtime": 100.7759,
+ "eval_samples_per_second": 14.091,
+ "eval_steps_per_second": 1.766,
+ "step": 3760
+ },
+ {
+ "entropy": 0.09276993470266462,
+ "epoch": 9.108564535585042,
+ "grad_norm": 0.0987873300909996,
+ "learning_rate": 1.0440616602241303e-05,
+ "loss": 0.06252194046974183,
+ "mean_token_accuracy": 0.9791185334324837,
+ "num_tokens": 9217543.0,
+ "step": 3780
+ },
+ {
+ "epoch": 9.108564535585042,
+ "eval_entropy": 0.18163840397355263,
+ "eval_loss": 1.124428629875183,
+ "eval_mean_token_accuracy": 0.8379190537366974,
+ "eval_num_tokens": 9217543.0,
+ "eval_runtime": 100.7229,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3780
+ },
+ {
+ "entropy": 0.08816529829055071,
+ "epoch": 9.156815440289506,
+ "grad_norm": 0.10476374626159668,
+ "learning_rate": 9.353256943939065e-06,
+ "loss": 0.056501579284667966,
+ "mean_token_accuracy": 0.9800700470805168,
+ "num_tokens": 9268724.0,
+ "step": 3800
+ },
+ {
+ "epoch": 9.156815440289506,
+ "eval_entropy": 0.17974376218037658,
+ "eval_loss": 1.1325938701629639,
+ "eval_mean_token_accuracy": 0.8377991332097,
+ "eval_num_tokens": 9268724.0,
+ "eval_runtime": 100.8729,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3800
+ },
+ {
+ "entropy": 0.0886100891046226,
+ "epoch": 9.205066344993968,
+ "grad_norm": 0.10369760543107986,
+ "learning_rate": 8.32442700674502e-06,
+ "loss": 0.05948063135147095,
+ "mean_token_accuracy": 0.9798692613840103,
+ "num_tokens": 9316805.0,
+ "step": 3820
+ },
+ {
+ "epoch": 9.205066344993968,
+ "eval_entropy": 0.1793072780364015,
+ "eval_loss": 1.1363310813903809,
+ "eval_mean_token_accuracy": 0.8378076965219519,
+ "eval_num_tokens": 9316805.0,
+ "eval_runtime": 100.8909,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3820
+ },
+ {
+ "entropy": 0.07910045385360717,
+ "epoch": 9.253317249698432,
+ "grad_norm": 0.06480544060468674,
+ "learning_rate": 7.3544179373615505e-06,
+ "loss": 0.053755837678909305,
+ "mean_token_accuracy": 0.9819905251264572,
+ "num_tokens": 9369940.0,
+ "step": 3840
+ },
+ {
+ "epoch": 9.253317249698432,
+ "eval_entropy": 0.1782708638086078,
+ "eval_loss": 1.1404398679733276,
+ "eval_mean_token_accuracy": 0.8379109310969878,
+ "eval_num_tokens": 9369940.0,
+ "eval_runtime": 100.7264,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3840
+ },
+ {
+ "entropy": 0.09386134278029204,
+ "epoch": 9.301568154402895,
+ "grad_norm": 0.12832233309745789,
+ "learning_rate": 6.443504236881163e-06,
+ "loss": 0.061395853757858276,
+ "mean_token_accuracy": 0.9780228048563003,
+ "num_tokens": 9416788.0,
+ "step": 3860
+ },
+ {
+ "epoch": 9.301568154402895,
+ "eval_entropy": 0.1777799724276816,
+ "eval_loss": 1.1431900262832642,
+ "eval_mean_token_accuracy": 0.8377661607908399,
+ "eval_num_tokens": 9416788.0,
+ "eval_runtime": 100.8637,
+ "eval_samples_per_second": 14.078,
+ "eval_steps_per_second": 1.765,
+ "step": 3860
+ },
+ {
+ "entropy": 0.0827090598642826,
+ "epoch": 9.349819059107359,
+ "grad_norm": 0.12366542220115662,
+ "learning_rate": 5.591943683106039e-06,
+ "loss": 0.05561224818229675,
+ "mean_token_accuracy": 0.9803308844566345,
+ "num_tokens": 9467862.0,
+ "step": 3880
+ },
+ {
+ "epoch": 9.349819059107359,
+ "eval_entropy": 0.1778556579368168,
+ "eval_loss": 1.1441748142242432,
+ "eval_mean_token_accuracy": 0.8377985549107027,
+ "eval_num_tokens": 9467862.0,
+ "eval_runtime": 100.7182,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3880
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.963460779994112e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3900/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3900/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3900/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3900/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3900/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3900/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3900/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3900/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3900/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3900/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3900/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3900/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3900/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3900/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..df690e45d5bc0f802dcb647486d643b4dc79c832
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3900/trainer_state.json
@@ -0,0 +1,4129 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 9.398069963811821,
+ "eval_steps": 20,
+ "global_step": 3900,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ },
+ {
+ "entropy": 0.0912214593961835,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.16693317890167236,
+ "learning_rate": 1.6744016157207948e-05,
+ "loss": 0.06410449743270874,
+ "mean_token_accuracy": 0.9774021029472351,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.18701685898089676,
+ "eval_loss": 1.1080079078674316,
+ "eval_mean_token_accuracy": 0.8379885400279184,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 100.8862,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3680
+ },
+ {
+ "entropy": 0.09622923573479056,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.15017707645893097,
+ "learning_rate": 1.53689516610855e-05,
+ "loss": 0.06397625803947449,
+ "mean_token_accuracy": 0.9777478039264679,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.1863192869873529,
+ "eval_loss": 1.1090794801712036,
+ "eval_mean_token_accuracy": 0.8382032892007506,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 100.8767,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3700
+ },
+ {
+ "entropy": 0.0912615241482854,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.1881953328847885,
+ "learning_rate": 1.4050714515625077e-05,
+ "loss": 0.062059921026229856,
+ "mean_token_accuracy": 0.9778110668063164,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.18606430442815416,
+ "eval_loss": 1.109899640083313,
+ "eval_mean_token_accuracy": 0.8379705333977603,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 100.7437,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3720
+ },
+ {
+ "entropy": 0.09705390800268222,
+ "epoch": 9.012062726176115,
+ "grad_norm": 0.3272952437400818,
+ "learning_rate": 1.2789677766345677e-05,
+ "loss": 0.06879411935806275,
+ "mean_token_accuracy": 0.9774823463880099,
+ "num_tokens": 9122597.0,
+ "step": 3740
+ },
+ {
+ "epoch": 9.012062726176115,
+ "eval_entropy": 0.18716304565078756,
+ "eval_loss": 1.103650689125061,
+ "eval_mean_token_accuracy": 0.8379069282097763,
+ "eval_num_tokens": 9122597.0,
+ "eval_runtime": 100.8591,
+ "eval_samples_per_second": 14.079,
+ "eval_steps_per_second": 1.765,
+ "step": 3740
+ },
+ {
+ "entropy": 0.09239197438582777,
+ "epoch": 9.060313630880579,
+ "grad_norm": 0.12848101556301117,
+ "learning_rate": 1.1586198271730714e-05,
+ "loss": 0.058521777391433716,
+ "mean_token_accuracy": 0.9808280184864998,
+ "num_tokens": 9171477.0,
+ "step": 3760
+ },
+ {
+ "epoch": 9.060313630880579,
+ "eval_entropy": 0.18443025656965342,
+ "eval_loss": 1.1128716468811035,
+ "eval_mean_token_accuracy": 0.8376701189560837,
+ "eval_num_tokens": 9171477.0,
+ "eval_runtime": 100.7759,
+ "eval_samples_per_second": 14.091,
+ "eval_steps_per_second": 1.766,
+ "step": 3760
+ },
+ {
+ "entropy": 0.09276993470266462,
+ "epoch": 9.108564535585042,
+ "grad_norm": 0.0987873300909996,
+ "learning_rate": 1.0440616602241303e-05,
+ "loss": 0.06252194046974183,
+ "mean_token_accuracy": 0.9791185334324837,
+ "num_tokens": 9217543.0,
+ "step": 3780
+ },
+ {
+ "epoch": 9.108564535585042,
+ "eval_entropy": 0.18163840397355263,
+ "eval_loss": 1.124428629875183,
+ "eval_mean_token_accuracy": 0.8379190537366974,
+ "eval_num_tokens": 9217543.0,
+ "eval_runtime": 100.7229,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3780
+ },
+ {
+ "entropy": 0.08816529829055071,
+ "epoch": 9.156815440289506,
+ "grad_norm": 0.10476374626159668,
+ "learning_rate": 9.353256943939065e-06,
+ "loss": 0.056501579284667966,
+ "mean_token_accuracy": 0.9800700470805168,
+ "num_tokens": 9268724.0,
+ "step": 3800
+ },
+ {
+ "epoch": 9.156815440289506,
+ "eval_entropy": 0.17974376218037658,
+ "eval_loss": 1.1325938701629639,
+ "eval_mean_token_accuracy": 0.8377991332097,
+ "eval_num_tokens": 9268724.0,
+ "eval_runtime": 100.8729,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3800
+ },
+ {
+ "entropy": 0.0886100891046226,
+ "epoch": 9.205066344993968,
+ "grad_norm": 0.10369760543107986,
+ "learning_rate": 8.32442700674502e-06,
+ "loss": 0.05948063135147095,
+ "mean_token_accuracy": 0.9798692613840103,
+ "num_tokens": 9316805.0,
+ "step": 3820
+ },
+ {
+ "epoch": 9.205066344993968,
+ "eval_entropy": 0.1793072780364015,
+ "eval_loss": 1.1363310813903809,
+ "eval_mean_token_accuracy": 0.8378076965219519,
+ "eval_num_tokens": 9316805.0,
+ "eval_runtime": 100.8909,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3820
+ },
+ {
+ "entropy": 0.07910045385360717,
+ "epoch": 9.253317249698432,
+ "grad_norm": 0.06480544060468674,
+ "learning_rate": 7.3544179373615505e-06,
+ "loss": 0.053755837678909305,
+ "mean_token_accuracy": 0.9819905251264572,
+ "num_tokens": 9369940.0,
+ "step": 3840
+ },
+ {
+ "epoch": 9.253317249698432,
+ "eval_entropy": 0.1782708638086078,
+ "eval_loss": 1.1404398679733276,
+ "eval_mean_token_accuracy": 0.8379109310969878,
+ "eval_num_tokens": 9369940.0,
+ "eval_runtime": 100.7264,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3840
+ },
+ {
+ "entropy": 0.09386134278029204,
+ "epoch": 9.301568154402895,
+ "grad_norm": 0.12832233309745789,
+ "learning_rate": 6.443504236881163e-06,
+ "loss": 0.061395853757858276,
+ "mean_token_accuracy": 0.9780228048563003,
+ "num_tokens": 9416788.0,
+ "step": 3860
+ },
+ {
+ "epoch": 9.301568154402895,
+ "eval_entropy": 0.1777799724276816,
+ "eval_loss": 1.1431900262832642,
+ "eval_mean_token_accuracy": 0.8377661607908399,
+ "eval_num_tokens": 9416788.0,
+ "eval_runtime": 100.8637,
+ "eval_samples_per_second": 14.078,
+ "eval_steps_per_second": 1.765,
+ "step": 3860
+ },
+ {
+ "entropy": 0.0827090598642826,
+ "epoch": 9.349819059107359,
+ "grad_norm": 0.12366542220115662,
+ "learning_rate": 5.591943683106039e-06,
+ "loss": 0.05561224818229675,
+ "mean_token_accuracy": 0.9803308844566345,
+ "num_tokens": 9467862.0,
+ "step": 3880
+ },
+ {
+ "epoch": 9.349819059107359,
+ "eval_entropy": 0.1778556579368168,
+ "eval_loss": 1.1441748142242432,
+ "eval_mean_token_accuracy": 0.8377985549107027,
+ "eval_num_tokens": 9467862.0,
+ "eval_runtime": 100.7182,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3880
+ },
+ {
+ "entropy": 0.09148634811863303,
+ "epoch": 9.398069963811821,
+ "grad_norm": 0.1590287685394287,
+ "learning_rate": 4.799977257599879e-06,
+ "loss": 0.06154900789260864,
+ "mean_token_accuracy": 0.9787917152047158,
+ "num_tokens": 9514611.0,
+ "step": 3900
+ },
+ {
+ "epoch": 9.398069963811821,
+ "eval_entropy": 0.17795597113082917,
+ "eval_loss": 1.1449599266052246,
+ "eval_mean_token_accuracy": 0.8378339922160245,
+ "eval_num_tokens": 9514611.0,
+ "eval_runtime": 100.8791,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3900
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.97314848536064e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3920/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3920/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3920/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3920/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3920/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3920/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3920/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3920/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3920/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3920/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3920/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3920/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3920/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3920/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..2c850514b30c8572da738b12f0f7fe7d2bc039f6
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3920/trainer_state.json
@@ -0,0 +1,4150 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 9.446320868516285,
+ "eval_steps": 20,
+ "global_step": 3920,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ },
+ {
+ "entropy": 0.0912214593961835,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.16693317890167236,
+ "learning_rate": 1.6744016157207948e-05,
+ "loss": 0.06410449743270874,
+ "mean_token_accuracy": 0.9774021029472351,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.18701685898089676,
+ "eval_loss": 1.1080079078674316,
+ "eval_mean_token_accuracy": 0.8379885400279184,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 100.8862,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3680
+ },
+ {
+ "entropy": 0.09622923573479056,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.15017707645893097,
+ "learning_rate": 1.53689516610855e-05,
+ "loss": 0.06397625803947449,
+ "mean_token_accuracy": 0.9777478039264679,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.1863192869873529,
+ "eval_loss": 1.1090794801712036,
+ "eval_mean_token_accuracy": 0.8382032892007506,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 100.8767,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3700
+ },
+ {
+ "entropy": 0.0912615241482854,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.1881953328847885,
+ "learning_rate": 1.4050714515625077e-05,
+ "loss": 0.062059921026229856,
+ "mean_token_accuracy": 0.9778110668063164,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.18606430442815416,
+ "eval_loss": 1.109899640083313,
+ "eval_mean_token_accuracy": 0.8379705333977603,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 100.7437,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3720
+ },
+ {
+ "entropy": 0.09705390800268222,
+ "epoch": 9.012062726176115,
+ "grad_norm": 0.3272952437400818,
+ "learning_rate": 1.2789677766345677e-05,
+ "loss": 0.06879411935806275,
+ "mean_token_accuracy": 0.9774823463880099,
+ "num_tokens": 9122597.0,
+ "step": 3740
+ },
+ {
+ "epoch": 9.012062726176115,
+ "eval_entropy": 0.18716304565078756,
+ "eval_loss": 1.103650689125061,
+ "eval_mean_token_accuracy": 0.8379069282097763,
+ "eval_num_tokens": 9122597.0,
+ "eval_runtime": 100.8591,
+ "eval_samples_per_second": 14.079,
+ "eval_steps_per_second": 1.765,
+ "step": 3740
+ },
+ {
+ "entropy": 0.09239197438582777,
+ "epoch": 9.060313630880579,
+ "grad_norm": 0.12848101556301117,
+ "learning_rate": 1.1586198271730714e-05,
+ "loss": 0.058521777391433716,
+ "mean_token_accuracy": 0.9808280184864998,
+ "num_tokens": 9171477.0,
+ "step": 3760
+ },
+ {
+ "epoch": 9.060313630880579,
+ "eval_entropy": 0.18443025656965342,
+ "eval_loss": 1.1128716468811035,
+ "eval_mean_token_accuracy": 0.8376701189560837,
+ "eval_num_tokens": 9171477.0,
+ "eval_runtime": 100.7759,
+ "eval_samples_per_second": 14.091,
+ "eval_steps_per_second": 1.766,
+ "step": 3760
+ },
+ {
+ "entropy": 0.09276993470266462,
+ "epoch": 9.108564535585042,
+ "grad_norm": 0.0987873300909996,
+ "learning_rate": 1.0440616602241303e-05,
+ "loss": 0.06252194046974183,
+ "mean_token_accuracy": 0.9791185334324837,
+ "num_tokens": 9217543.0,
+ "step": 3780
+ },
+ {
+ "epoch": 9.108564535585042,
+ "eval_entropy": 0.18163840397355263,
+ "eval_loss": 1.124428629875183,
+ "eval_mean_token_accuracy": 0.8379190537366974,
+ "eval_num_tokens": 9217543.0,
+ "eval_runtime": 100.7229,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3780
+ },
+ {
+ "entropy": 0.08816529829055071,
+ "epoch": 9.156815440289506,
+ "grad_norm": 0.10476374626159668,
+ "learning_rate": 9.353256943939065e-06,
+ "loss": 0.056501579284667966,
+ "mean_token_accuracy": 0.9800700470805168,
+ "num_tokens": 9268724.0,
+ "step": 3800
+ },
+ {
+ "epoch": 9.156815440289506,
+ "eval_entropy": 0.17974376218037658,
+ "eval_loss": 1.1325938701629639,
+ "eval_mean_token_accuracy": 0.8377991332097,
+ "eval_num_tokens": 9268724.0,
+ "eval_runtime": 100.8729,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3800
+ },
+ {
+ "entropy": 0.0886100891046226,
+ "epoch": 9.205066344993968,
+ "grad_norm": 0.10369760543107986,
+ "learning_rate": 8.32442700674502e-06,
+ "loss": 0.05948063135147095,
+ "mean_token_accuracy": 0.9798692613840103,
+ "num_tokens": 9316805.0,
+ "step": 3820
+ },
+ {
+ "epoch": 9.205066344993968,
+ "eval_entropy": 0.1793072780364015,
+ "eval_loss": 1.1363310813903809,
+ "eval_mean_token_accuracy": 0.8378076965219519,
+ "eval_num_tokens": 9316805.0,
+ "eval_runtime": 100.8909,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3820
+ },
+ {
+ "entropy": 0.07910045385360717,
+ "epoch": 9.253317249698432,
+ "grad_norm": 0.06480544060468674,
+ "learning_rate": 7.3544179373615505e-06,
+ "loss": 0.053755837678909305,
+ "mean_token_accuracy": 0.9819905251264572,
+ "num_tokens": 9369940.0,
+ "step": 3840
+ },
+ {
+ "epoch": 9.253317249698432,
+ "eval_entropy": 0.1782708638086078,
+ "eval_loss": 1.1404398679733276,
+ "eval_mean_token_accuracy": 0.8379109310969878,
+ "eval_num_tokens": 9369940.0,
+ "eval_runtime": 100.7264,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3840
+ },
+ {
+ "entropy": 0.09386134278029204,
+ "epoch": 9.301568154402895,
+ "grad_norm": 0.12832233309745789,
+ "learning_rate": 6.443504236881163e-06,
+ "loss": 0.061395853757858276,
+ "mean_token_accuracy": 0.9780228048563003,
+ "num_tokens": 9416788.0,
+ "step": 3860
+ },
+ {
+ "epoch": 9.301568154402895,
+ "eval_entropy": 0.1777799724276816,
+ "eval_loss": 1.1431900262832642,
+ "eval_mean_token_accuracy": 0.8377661607908399,
+ "eval_num_tokens": 9416788.0,
+ "eval_runtime": 100.8637,
+ "eval_samples_per_second": 14.078,
+ "eval_steps_per_second": 1.765,
+ "step": 3860
+ },
+ {
+ "entropy": 0.0827090598642826,
+ "epoch": 9.349819059107359,
+ "grad_norm": 0.12366542220115662,
+ "learning_rate": 5.591943683106039e-06,
+ "loss": 0.05561224818229675,
+ "mean_token_accuracy": 0.9803308844566345,
+ "num_tokens": 9467862.0,
+ "step": 3880
+ },
+ {
+ "epoch": 9.349819059107359,
+ "eval_entropy": 0.1778556579368168,
+ "eval_loss": 1.1441748142242432,
+ "eval_mean_token_accuracy": 0.8377985549107027,
+ "eval_num_tokens": 9467862.0,
+ "eval_runtime": 100.7182,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3880
+ },
+ {
+ "entropy": 0.09148634811863303,
+ "epoch": 9.398069963811821,
+ "grad_norm": 0.1590287685394287,
+ "learning_rate": 4.799977257599879e-06,
+ "loss": 0.06154900789260864,
+ "mean_token_accuracy": 0.9787917152047158,
+ "num_tokens": 9514611.0,
+ "step": 3900
+ },
+ {
+ "epoch": 9.398069963811821,
+ "eval_entropy": 0.17795597113082917,
+ "eval_loss": 1.1449599266052246,
+ "eval_mean_token_accuracy": 0.8378339922160245,
+ "eval_num_tokens": 9514611.0,
+ "eval_runtime": 100.8791,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3900
+ },
+ {
+ "entropy": 0.0879560124129057,
+ "epoch": 9.446320868516285,
+ "grad_norm": 0.11894718557596207,
+ "learning_rate": 4.067829077493076e-06,
+ "loss": 0.05804988145828247,
+ "mean_token_accuracy": 0.9796567305922508,
+ "num_tokens": 9563437.0,
+ "step": 3920
+ },
+ {
+ "epoch": 9.446320868516285,
+ "eval_entropy": 0.17797360205081073,
+ "eval_loss": 1.146341323852539,
+ "eval_mean_token_accuracy": 0.8379327249660921,
+ "eval_num_tokens": 9563437.0,
+ "eval_runtime": 100.8092,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3920
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.983131622657024e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3940/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3940/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3940/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3940/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3940/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3940/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3940/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3940/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3940/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3940/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3940/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3940/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3940/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3940/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..41ef0365d7293abc8a35c4b9b1675460a7a6e5db
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3940/trainer_state.json
@@ -0,0 +1,4171 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 9.494571773220748,
+ "eval_steps": 20,
+ "global_step": 3940,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ },
+ {
+ "entropy": 0.0912214593961835,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.16693317890167236,
+ "learning_rate": 1.6744016157207948e-05,
+ "loss": 0.06410449743270874,
+ "mean_token_accuracy": 0.9774021029472351,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.18701685898089676,
+ "eval_loss": 1.1080079078674316,
+ "eval_mean_token_accuracy": 0.8379885400279184,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 100.8862,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3680
+ },
+ {
+ "entropy": 0.09622923573479056,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.15017707645893097,
+ "learning_rate": 1.53689516610855e-05,
+ "loss": 0.06397625803947449,
+ "mean_token_accuracy": 0.9777478039264679,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.1863192869873529,
+ "eval_loss": 1.1090794801712036,
+ "eval_mean_token_accuracy": 0.8382032892007506,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 100.8767,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3700
+ },
+ {
+ "entropy": 0.0912615241482854,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.1881953328847885,
+ "learning_rate": 1.4050714515625077e-05,
+ "loss": 0.062059921026229856,
+ "mean_token_accuracy": 0.9778110668063164,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.18606430442815416,
+ "eval_loss": 1.109899640083313,
+ "eval_mean_token_accuracy": 0.8379705333977603,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 100.7437,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3720
+ },
+ {
+ "entropy": 0.09705390800268222,
+ "epoch": 9.012062726176115,
+ "grad_norm": 0.3272952437400818,
+ "learning_rate": 1.2789677766345677e-05,
+ "loss": 0.06879411935806275,
+ "mean_token_accuracy": 0.9774823463880099,
+ "num_tokens": 9122597.0,
+ "step": 3740
+ },
+ {
+ "epoch": 9.012062726176115,
+ "eval_entropy": 0.18716304565078756,
+ "eval_loss": 1.103650689125061,
+ "eval_mean_token_accuracy": 0.8379069282097763,
+ "eval_num_tokens": 9122597.0,
+ "eval_runtime": 100.8591,
+ "eval_samples_per_second": 14.079,
+ "eval_steps_per_second": 1.765,
+ "step": 3740
+ },
+ {
+ "entropy": 0.09239197438582777,
+ "epoch": 9.060313630880579,
+ "grad_norm": 0.12848101556301117,
+ "learning_rate": 1.1586198271730714e-05,
+ "loss": 0.058521777391433716,
+ "mean_token_accuracy": 0.9808280184864998,
+ "num_tokens": 9171477.0,
+ "step": 3760
+ },
+ {
+ "epoch": 9.060313630880579,
+ "eval_entropy": 0.18443025656965342,
+ "eval_loss": 1.1128716468811035,
+ "eval_mean_token_accuracy": 0.8376701189560837,
+ "eval_num_tokens": 9171477.0,
+ "eval_runtime": 100.7759,
+ "eval_samples_per_second": 14.091,
+ "eval_steps_per_second": 1.766,
+ "step": 3760
+ },
+ {
+ "entropy": 0.09276993470266462,
+ "epoch": 9.108564535585042,
+ "grad_norm": 0.0987873300909996,
+ "learning_rate": 1.0440616602241303e-05,
+ "loss": 0.06252194046974183,
+ "mean_token_accuracy": 0.9791185334324837,
+ "num_tokens": 9217543.0,
+ "step": 3780
+ },
+ {
+ "epoch": 9.108564535585042,
+ "eval_entropy": 0.18163840397355263,
+ "eval_loss": 1.124428629875183,
+ "eval_mean_token_accuracy": 0.8379190537366974,
+ "eval_num_tokens": 9217543.0,
+ "eval_runtime": 100.7229,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3780
+ },
+ {
+ "entropy": 0.08816529829055071,
+ "epoch": 9.156815440289506,
+ "grad_norm": 0.10476374626159668,
+ "learning_rate": 9.353256943939065e-06,
+ "loss": 0.056501579284667966,
+ "mean_token_accuracy": 0.9800700470805168,
+ "num_tokens": 9268724.0,
+ "step": 3800
+ },
+ {
+ "epoch": 9.156815440289506,
+ "eval_entropy": 0.17974376218037658,
+ "eval_loss": 1.1325938701629639,
+ "eval_mean_token_accuracy": 0.8377991332097,
+ "eval_num_tokens": 9268724.0,
+ "eval_runtime": 100.8729,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3800
+ },
+ {
+ "entropy": 0.0886100891046226,
+ "epoch": 9.205066344993968,
+ "grad_norm": 0.10369760543107986,
+ "learning_rate": 8.32442700674502e-06,
+ "loss": 0.05948063135147095,
+ "mean_token_accuracy": 0.9798692613840103,
+ "num_tokens": 9316805.0,
+ "step": 3820
+ },
+ {
+ "epoch": 9.205066344993968,
+ "eval_entropy": 0.1793072780364015,
+ "eval_loss": 1.1363310813903809,
+ "eval_mean_token_accuracy": 0.8378076965219519,
+ "eval_num_tokens": 9316805.0,
+ "eval_runtime": 100.8909,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3820
+ },
+ {
+ "entropy": 0.07910045385360717,
+ "epoch": 9.253317249698432,
+ "grad_norm": 0.06480544060468674,
+ "learning_rate": 7.3544179373615505e-06,
+ "loss": 0.053755837678909305,
+ "mean_token_accuracy": 0.9819905251264572,
+ "num_tokens": 9369940.0,
+ "step": 3840
+ },
+ {
+ "epoch": 9.253317249698432,
+ "eval_entropy": 0.1782708638086078,
+ "eval_loss": 1.1404398679733276,
+ "eval_mean_token_accuracy": 0.8379109310969878,
+ "eval_num_tokens": 9369940.0,
+ "eval_runtime": 100.7264,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3840
+ },
+ {
+ "entropy": 0.09386134278029204,
+ "epoch": 9.301568154402895,
+ "grad_norm": 0.12832233309745789,
+ "learning_rate": 6.443504236881163e-06,
+ "loss": 0.061395853757858276,
+ "mean_token_accuracy": 0.9780228048563003,
+ "num_tokens": 9416788.0,
+ "step": 3860
+ },
+ {
+ "epoch": 9.301568154402895,
+ "eval_entropy": 0.1777799724276816,
+ "eval_loss": 1.1431900262832642,
+ "eval_mean_token_accuracy": 0.8377661607908399,
+ "eval_num_tokens": 9416788.0,
+ "eval_runtime": 100.8637,
+ "eval_samples_per_second": 14.078,
+ "eval_steps_per_second": 1.765,
+ "step": 3860
+ },
+ {
+ "entropy": 0.0827090598642826,
+ "epoch": 9.349819059107359,
+ "grad_norm": 0.12366542220115662,
+ "learning_rate": 5.591943683106039e-06,
+ "loss": 0.05561224818229675,
+ "mean_token_accuracy": 0.9803308844566345,
+ "num_tokens": 9467862.0,
+ "step": 3880
+ },
+ {
+ "epoch": 9.349819059107359,
+ "eval_entropy": 0.1778556579368168,
+ "eval_loss": 1.1441748142242432,
+ "eval_mean_token_accuracy": 0.8377985549107027,
+ "eval_num_tokens": 9467862.0,
+ "eval_runtime": 100.7182,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3880
+ },
+ {
+ "entropy": 0.09148634811863303,
+ "epoch": 9.398069963811821,
+ "grad_norm": 0.1590287685394287,
+ "learning_rate": 4.799977257599879e-06,
+ "loss": 0.06154900789260864,
+ "mean_token_accuracy": 0.9787917152047158,
+ "num_tokens": 9514611.0,
+ "step": 3900
+ },
+ {
+ "epoch": 9.398069963811821,
+ "eval_entropy": 0.17795597113082917,
+ "eval_loss": 1.1449599266052246,
+ "eval_mean_token_accuracy": 0.8378339922160245,
+ "eval_num_tokens": 9514611.0,
+ "eval_runtime": 100.8791,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3900
+ },
+ {
+ "entropy": 0.0879560124129057,
+ "epoch": 9.446320868516285,
+ "grad_norm": 0.11894718557596207,
+ "learning_rate": 4.067829077493076e-06,
+ "loss": 0.05804988145828247,
+ "mean_token_accuracy": 0.9796567305922508,
+ "num_tokens": 9563437.0,
+ "step": 3920
+ },
+ {
+ "epoch": 9.446320868516285,
+ "eval_entropy": 0.17797360205081073,
+ "eval_loss": 1.146341323852539,
+ "eval_mean_token_accuracy": 0.8379327249660921,
+ "eval_num_tokens": 9563437.0,
+ "eval_runtime": 100.8092,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3920
+ },
+ {
+ "entropy": 0.0873468822799623,
+ "epoch": 9.494571773220748,
+ "grad_norm": 0.09161137789487839,
+ "learning_rate": 3.3957063320601047e-06,
+ "loss": 0.05694714784622192,
+ "mean_token_accuracy": 0.9784631252288818,
+ "num_tokens": 9614034.0,
+ "step": 3940
+ },
+ {
+ "epoch": 9.494571773220748,
+ "eval_entropy": 0.17779575903596503,
+ "eval_loss": 1.148206114768982,
+ "eval_mean_token_accuracy": 0.8378150037835154,
+ "eval_num_tokens": 9614034.0,
+ "eval_runtime": 100.7077,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3940
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 1.99420484906496e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3960/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3960/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3960/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3960/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3960/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..b359bb4729fdaf85a25d259c1ba2fc28433b2a4a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3960/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 128,
+ "lora_bias": false,
+ "lora_dropout": 0.0539675908752296,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3960/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3960/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3960/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3960/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3960/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3960/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3960/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3960/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..519e920c2da0c996f9c3c163e6e52ddef84eb83d
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test1/checkpoint-3960/trainer_state.json
@@ -0,0 +1,4192 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 9.54282267792521,
+ "eval_steps": 20,
+ "global_step": 3960,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8160341799259185,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 1.1313843727111816,
+ "learning_rate": 1.9794859683580766e-05,
+ "loss": 1.7627403259277343,
+ "mean_token_accuracy": 0.6354069545865059,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.538598193881217,
+ "eval_loss": 1.4881398677825928,
+ "eval_mean_token_accuracy": 0.686459881201219,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 100.8923,
+ "eval_samples_per_second": 14.074,
+ "eval_steps_per_second": 1.764,
+ "step": 20
+ },
+ {
+ "entropy": 1.1673978716135025,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.6167025566101074,
+ "learning_rate": 4.0631554087349995e-05,
+ "loss": 1.0887075424194337,
+ "mean_token_accuracy": 0.737513878941536,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 0.887577452351538,
+ "eval_loss": 0.8519237041473389,
+ "eval_mean_token_accuracy": 0.7741017204322173,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 100.4819,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 40
+ },
+ {
+ "entropy": 0.8220535531640053,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.8396430611610413,
+ "learning_rate": 6.146824849111923e-05,
+ "loss": 0.7812146186828614,
+ "mean_token_accuracy": 0.7872247830033302,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 0.770665143982748,
+ "eval_loss": 0.7478981614112854,
+ "eval_mean_token_accuracy": 0.7949167068084974,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 100.6518,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 60
+ },
+ {
+ "entropy": 0.7776695728302002,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.5731202363967896,
+ "learning_rate": 8.230494289488845e-05,
+ "loss": 0.732721471786499,
+ "mean_token_accuracy": 0.7976997002959252,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 0.7650937618834249,
+ "eval_loss": 0.7109046578407288,
+ "eval_mean_token_accuracy": 0.8018777065732506,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 100.5983,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 80
+ },
+ {
+ "entropy": 0.7309209659695626,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.48046866059303284,
+ "learning_rate": 0.00010314163729865769,
+ "loss": 0.680583143234253,
+ "mean_token_accuracy": 0.8073364540934562,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.7152130687504672,
+ "eval_loss": 0.6989707350730896,
+ "eval_mean_token_accuracy": 0.8057264320636064,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 100.6264,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 100
+ },
+ {
+ "entropy": 0.7105271726846695,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.5512468814849854,
+ "learning_rate": 0.00012397833170242693,
+ "loss": 0.6674309253692627,
+ "mean_token_accuracy": 0.8098619893193245,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.7314498555124476,
+ "eval_loss": 0.6796280145645142,
+ "eval_mean_token_accuracy": 0.809063873264227,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 100.6132,
+ "eval_samples_per_second": 14.113,
+ "eval_steps_per_second": 1.769,
+ "step": 120
+ },
+ {
+ "entropy": 0.7079856500029564,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5903022885322571,
+ "learning_rate": 0.00014481502610619613,
+ "loss": 0.6623000144958496,
+ "mean_token_accuracy": 0.8101129323244095,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.7005330203624254,
+ "eval_loss": 0.671779453754425,
+ "eval_mean_token_accuracy": 0.8106091374091888,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 100.6904,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 140
+ },
+ {
+ "entropy": 0.7109143257141113,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.4516378343105316,
+ "learning_rate": 0.00016565172050996536,
+ "loss": 0.658915901184082,
+ "mean_token_accuracy": 0.8112782314419746,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.6891461650976974,
+ "eval_loss": 0.6598312258720398,
+ "eval_mean_token_accuracy": 0.8133360951134329,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 100.5057,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 160
+ },
+ {
+ "entropy": 0.6896152347326279,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.49479565024375916,
+ "learning_rate": 0.0001864884149137346,
+ "loss": 0.6466911315917969,
+ "mean_token_accuracy": 0.8141682580113411,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.6984815888860253,
+ "eval_loss": 0.652909517288208,
+ "eval_mean_token_accuracy": 0.8133517600847094,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 100.5867,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 180
+ },
+ {
+ "entropy": 0.6829591259360314,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4835337698459625,
+ "learning_rate": 0.00020732510931750382,
+ "loss": 0.6401876449584961,
+ "mean_token_accuracy": 0.8175502926111221,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.6963563973314306,
+ "eval_loss": 0.650087296962738,
+ "eval_mean_token_accuracy": 0.8159684401549651,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 100.6277,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 200
+ },
+ {
+ "entropy": 0.6903425320982933,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.44980454444885254,
+ "learning_rate": 0.00022816180372127305,
+ "loss": 0.6409444808959961,
+ "mean_token_accuracy": 0.8163257524371147,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.6687790182199371,
+ "eval_loss": 0.6451645493507385,
+ "eval_mean_token_accuracy": 0.8157875551266617,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 100.5444,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 220
+ },
+ {
+ "entropy": 0.6740772560238838,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.5312877297401428,
+ "learning_rate": 0.0002489984981250423,
+ "loss": 0.6307676315307618,
+ "mean_token_accuracy": 0.8174169659614563,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.7063750872451268,
+ "eval_loss": 0.6494986414909363,
+ "eval_mean_token_accuracy": 0.8156777198395032,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 100.5094,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 240
+ },
+ {
+ "entropy": 0.6876470118761062,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.46149370074272156,
+ "learning_rate": 0.00026983519252881153,
+ "loss": 0.6392203807830811,
+ "mean_token_accuracy": 0.8172321930527687,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.625594850838854,
+ "eval_loss": 0.6463819742202759,
+ "eval_mean_token_accuracy": 0.8179670525400826,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 100.5775,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 260
+ },
+ {
+ "entropy": 0.6673879995942116,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.42477288842201233,
+ "learning_rate": 0.0002906718869325807,
+ "loss": 0.6300291538238525,
+ "mean_token_accuracy": 0.8203139930963517,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.6863534494062488,
+ "eval_loss": 0.6392037272453308,
+ "eval_mean_token_accuracy": 0.8190520385677895,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 100.5106,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 280
+ },
+ {
+ "entropy": 0.6573980256915093,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.4656430780887604,
+ "learning_rate": 0.00031150858133635,
+ "loss": 0.6215964794158936,
+ "mean_token_accuracy": 0.8208559453487396,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.6834242812703165,
+ "eval_loss": 0.6479960083961487,
+ "eval_mean_token_accuracy": 0.8128752906001016,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 100.6101,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 300
+ },
+ {
+ "entropy": 0.6751310348510742,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.4495522081851959,
+ "learning_rate": 0.00033234527574011916,
+ "loss": 0.6402668952941895,
+ "mean_token_accuracy": 0.8170571058988572,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6920045178927733,
+ "eval_loss": 0.6402787566184998,
+ "eval_mean_token_accuracy": 0.817791668886549,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 100.6347,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 320
+ },
+ {
+ "entropy": 0.6846290081739426,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.5661432147026062,
+ "learning_rate": 0.00035318197014388845,
+ "loss": 0.6508871078491211,
+ "mean_token_accuracy": 0.8164169207215309,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6772527078564248,
+ "eval_loss": 0.6512064337730408,
+ "eval_mean_token_accuracy": 0.8170605137776793,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 100.5157,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 340
+ },
+ {
+ "entropy": 0.7022885024547577,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.6357236504554749,
+ "learning_rate": 0.0003740186645476577,
+ "loss": 0.6649135112762451,
+ "mean_token_accuracy": 0.8136772960424423,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.7435127543599418,
+ "eval_loss": 0.6573262214660645,
+ "eval_mean_token_accuracy": 0.8147802225659403,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 100.5271,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 360
+ },
+ {
+ "entropy": 0.6962979376316071,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.6179454326629639,
+ "learning_rate": 0.00039485535895142685,
+ "loss": 0.6509191513061523,
+ "mean_token_accuracy": 0.81364226937294,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6630894883295123,
+ "eval_loss": 0.6524088978767395,
+ "eval_mean_token_accuracy": 0.8165898587596551,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 100.5343,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 380
+ },
+ {
+ "entropy": 0.694319449365139,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.5458664894104004,
+ "learning_rate": 0.00041569205335519613,
+ "loss": 0.6624011993408203,
+ "mean_token_accuracy": 0.8136080905795098,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.7424925324622165,
+ "eval_loss": 0.6717661023139954,
+ "eval_mean_token_accuracy": 0.8099894781460922,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 100.5992,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 400
+ },
+ {
+ "entropy": 0.6861714736009256,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.6108037829399109,
+ "learning_rate": 0.0004323601853188671,
+ "loss": 0.6551628112792969,
+ "mean_token_accuracy": 0.8134882618219424,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6593745705786715,
+ "eval_loss": 0.6720747947692871,
+ "eval_mean_token_accuracy": 0.812802647941568,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 420
+ },
+ {
+ "entropy": 0.639520151913166,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.6054353713989258,
+ "learning_rate": 0.00043231736219609266,
+ "loss": 0.6096667766571044,
+ "mean_token_accuracy": 0.8227144703269005,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6316307876886946,
+ "eval_loss": 0.6621950268745422,
+ "eval_mean_token_accuracy": 0.8140057914712456,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 440
+ },
+ {
+ "entropy": 0.6544070363044738,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5111057758331299,
+ "learning_rate": 0.0004322133749558233,
+ "loss": 0.6176822185516357,
+ "mean_token_accuracy": 0.8201584875583648,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6714627766207363,
+ "eval_loss": 0.6738516092300415,
+ "eval_mean_token_accuracy": 0.8128616829936424,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 100.4986,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 460
+ },
+ {
+ "entropy": 0.6470697179436684,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.5434669256210327,
+ "learning_rate": 0.0004320482530252179,
+ "loss": 0.6144762516021729,
+ "mean_token_accuracy": 0.819926056265831,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6475969080844622,
+ "eval_loss": 0.6624214053153992,
+ "eval_mean_token_accuracy": 0.8137228820431098,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 100.5617,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 480
+ },
+ {
+ "entropy": 0.6659130036830903,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.6470707058906555,
+ "learning_rate": 0.0004318220431318296,
+ "loss": 0.6366308212280274,
+ "mean_token_accuracy": 0.8175261527299881,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.7232425132494295,
+ "eval_loss": 0.6877738833427429,
+ "eval_mean_token_accuracy": 0.8007767260744331,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 100.477,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 500
+ },
+ {
+ "entropy": 0.6759189993143082,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.6453561186790466,
+ "learning_rate": 0.00043153480929038245,
+ "loss": 0.637772512435913,
+ "mean_token_accuracy": 0.8162627041339874,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.7082584312792575,
+ "eval_loss": 0.6695764064788818,
+ "eval_mean_token_accuracy": 0.812019992410467,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 100.4957,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 520
+ },
+ {
+ "entropy": 0.6730152070522308,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.6234476566314697,
+ "learning_rate": 0.000431186632784656,
+ "loss": 0.640014362335205,
+ "mean_token_accuracy": 0.8168858170509339,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6655137846978862,
+ "eval_loss": 0.6578248739242554,
+ "eval_mean_token_accuracy": 0.8147600979617472,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 100.503,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 540
+ },
+ {
+ "entropy": 0.6636043325066566,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.5882512331008911,
+ "learning_rate": 0.0004307776121444827,
+ "loss": 0.6286775588989257,
+ "mean_token_accuracy": 0.8183174058794975,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6304813810948575,
+ "eval_loss": 0.6607436537742615,
+ "eval_mean_token_accuracy": 0.8169659268989992,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 100.6233,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 560
+ },
+ {
+ "entropy": 0.6608738481998444,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.6796398758888245,
+ "learning_rate": 0.00043030786311786564,
+ "loss": 0.622726583480835,
+ "mean_token_accuracy": 0.8230827242136002,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6645242689030894,
+ "eval_loss": 0.6662811636924744,
+ "eval_mean_token_accuracy": 0.8147623907314258,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 100.5074,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 580
+ },
+ {
+ "entropy": 0.6682895936071873,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5750718712806702,
+ "learning_rate": 0.0004297775186382226,
+ "loss": 0.6304276466369629,
+ "mean_token_accuracy": 0.8188625335693359,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6320532578765676,
+ "eval_loss": 0.650333046913147,
+ "eval_mean_token_accuracy": 0.8172853793321031,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 100.5116,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 600
+ },
+ {
+ "entropy": 0.653739221394062,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.5552534461021423,
+ "learning_rate": 0.00042918672878676804,
+ "loss": 0.6215689659118653,
+ "mean_token_accuracy": 0.8208896175026894,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6285932062382109,
+ "eval_loss": 0.6392536163330078,
+ "eval_mean_token_accuracy": 0.8205167825302381,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 100.6392,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 620
+ },
+ {
+ "entropy": 0.6476897612214089,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.536978542804718,
+ "learning_rate": 0.00042853566075004156,
+ "loss": 0.6131608009338378,
+ "mean_token_accuracy": 0.8208152890205384,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6626498856571283,
+ "eval_loss": 0.6479025483131409,
+ "eval_mean_token_accuracy": 0.8180894101603647,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 100.5035,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 640
+ },
+ {
+ "entropy": 0.6338633447885513,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.4518965780735016,
+ "learning_rate": 0.0004278244987725961,
+ "loss": 0.6010610580444335,
+ "mean_token_accuracy": 0.8247816175222397,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.64079723699709,
+ "eval_loss": 0.6417157053947449,
+ "eval_mean_token_accuracy": 0.8214994581897607,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 100.6393,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 660
+ },
+ {
+ "entropy": 0.6485912457108498,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.6028368473052979,
+ "learning_rate": 0.00042705344410485905,
+ "loss": 0.6120437145233154,
+ "mean_token_accuracy": 0.8236212745308876,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6863006427046958,
+ "eval_loss": 0.6420931816101074,
+ "eval_mean_token_accuracy": 0.8191533580924688,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 680
+ },
+ {
+ "entropy": 0.6459382772445679,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.5069310069084167,
+ "learning_rate": 0.00042622271494618073,
+ "loss": 0.6160889625549316,
+ "mean_token_accuracy": 0.8253828629851341,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6773919463157654,
+ "eval_loss": 0.6307896375656128,
+ "eval_mean_token_accuracy": 0.8217989825800563,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 100.6301,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 700
+ },
+ {
+ "entropy": 0.64687028080225,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.3830573856830597,
+ "learning_rate": 0.00042533254638308645,
+ "loss": 0.6104001998901367,
+ "mean_token_accuracy": 0.8255606815218925,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6179541145147902,
+ "eval_loss": 0.6170215606689453,
+ "eval_mean_token_accuracy": 0.8259220900160543,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 100.61,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 720
+ },
+ {
+ "entropy": 0.6306161753833294,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.5888957977294922,
+ "learning_rate": 0.0004243831903227498,
+ "loss": 0.5981546878814697,
+ "mean_token_accuracy": 0.8266776099801063,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6357416808605194,
+ "eval_loss": 0.6202242374420166,
+ "eval_mean_token_accuracy": 0.8253380304641937,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 100.5145,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 740
+ },
+ {
+ "entropy": 0.6290688961744308,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6138001680374146,
+ "learning_rate": 0.0004233749154217063,
+ "loss": 0.5853167057037354,
+ "mean_token_accuracy": 0.8281400203704834,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.6402715869833914,
+ "eval_loss": 0.6168262958526611,
+ "eval_mean_token_accuracy": 0.8246346306934785,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 100.5237,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 760
+ },
+ {
+ "entropy": 0.6218615680932998,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.574312686920166,
+ "learning_rate": 0.00042230800700982586,
+ "loss": 0.5841915607452393,
+ "mean_token_accuracy": 0.8304707780480385,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6301163148679091,
+ "eval_loss": 0.6073815226554871,
+ "eval_mean_token_accuracy": 0.8274695993809218,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 100.5111,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 780
+ },
+ {
+ "entropy": 0.6101033873856068,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.5054804682731628,
+ "learning_rate": 0.000421182767009569,
+ "loss": 0.5774452209472656,
+ "mean_token_accuracy": 0.8321943372488022,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.631400960215022,
+ "eval_loss": 0.6115217804908752,
+ "eval_mean_token_accuracy": 0.8249618762664581,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 100.4916,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 800
+ },
+ {
+ "entropy": 0.6217561915516854,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.5333649516105652,
+ "learning_rate": 0.0004199995138505454,
+ "loss": 0.5842787265777588,
+ "mean_token_accuracy": 0.8305525362491608,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.605928597657868,
+ "eval_loss": 0.6067516207695007,
+ "eval_mean_token_accuracy": 0.82715918408351,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 100.5286,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 820
+ },
+ {
+ "entropy": 0.5585076526189462,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.44819608330726624,
+ "learning_rate": 0.00041875858237940244,
+ "loss": 0.5252367496490479,
+ "mean_token_accuracy": 0.8405626966403081,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.6123150769579276,
+ "eval_loss": 0.6175946593284607,
+ "eval_mean_token_accuracy": 0.827637837509091,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 100.5414,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 840
+ },
+ {
+ "entropy": 0.5359115816652775,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.576781690120697,
+ "learning_rate": 0.0004174603237650678,
+ "loss": 0.5023207664489746,
+ "mean_token_accuracy": 0.8459693863987923,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5592134784111816,
+ "eval_loss": 0.620693564414978,
+ "eval_mean_token_accuracy": 0.827402507656076,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 100.6274,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 860
+ },
+ {
+ "entropy": 0.5214565828442573,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.45878875255584717,
+ "learning_rate": 0.00041610510539937234,
+ "loss": 0.48367843627929685,
+ "mean_token_accuracy": 0.8515541926026344,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.5860532344727034,
+ "eval_loss": 0.6134791970252991,
+ "eval_mean_token_accuracy": 0.8291774777214179,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 100.5295,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 880
+ },
+ {
+ "entropy": 0.5414379432797432,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5976294875144958,
+ "learning_rate": 0.0004146933107930829,
+ "loss": 0.5012908935546875,
+ "mean_token_accuracy": 0.8466357633471489,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5427783280611038,
+ "eval_loss": 0.612716019153595,
+ "eval_mean_token_accuracy": 0.829505049445656,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 900
+ },
+ {
+ "entropy": 0.521108990907669,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.5145775675773621,
+ "learning_rate": 0.000413225339467373,
+ "loss": 0.4877732276916504,
+ "mean_token_accuracy": 0.8504189744591713,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.5620684869838565,
+ "eval_loss": 0.6176337599754333,
+ "eval_mean_token_accuracy": 0.8291444483767735,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 100.6819,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 920
+ },
+ {
+ "entropy": 0.5452766507863999,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.4820452928543091,
+ "learning_rate": 0.0004117016068407636,
+ "loss": 0.4991307735443115,
+ "mean_token_accuracy": 0.8474363312125206,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5611683193887218,
+ "eval_loss": 0.6153580546379089,
+ "eval_mean_token_accuracy": 0.8286041272490212,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 100.5249,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 940
+ },
+ {
+ "entropy": 0.5572576269507408,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.3947993814945221,
+ "learning_rate": 0.00041012254411156387,
+ "loss": 0.5191003799438476,
+ "mean_token_accuracy": 0.8422838419675827,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5651392345682958,
+ "eval_loss": 0.6121005415916443,
+ "eval_mean_token_accuracy": 0.8286536242185014,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 100.7033,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 960
+ },
+ {
+ "entropy": 0.5441671140491963,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.7719791531562805,
+ "learning_rate": 0.0004084885981358477,
+ "loss": 0.5072606086730957,
+ "mean_token_accuracy": 0.8478495761752128,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5651004270891126,
+ "eval_loss": 0.6069501042366028,
+ "eval_mean_token_accuracy": 0.8311042156112328,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 100.5487,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 980
+ },
+ {
+ "entropy": 0.5462074346840382,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.5694436430931091,
+ "learning_rate": 0.00040680023130099856,
+ "loss": 0.5097150325775146,
+ "mean_token_accuracy": 0.8474388793110847,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5895893243926295,
+ "eval_loss": 0.6078168749809265,
+ "eval_mean_token_accuracy": 0.828030709470256,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 100.5413,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5473216332495212,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3989660143852234,
+ "learning_rate": 0.00040505792139485913,
+ "loss": 0.5087806224822998,
+ "mean_token_accuracy": 0.8473173275589942,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5588909702354603,
+ "eval_loss": 0.6087605953216553,
+ "eval_mean_token_accuracy": 0.8309365524334854,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 100.5137,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5430163390934467,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.4916608929634094,
+ "learning_rate": 0.00040326216147052336,
+ "loss": 0.5027707576751709,
+ "mean_token_accuracy": 0.8466959804296493,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5819033988741007,
+ "eval_loss": 0.609398365020752,
+ "eval_mean_token_accuracy": 0.8297666612635838,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 100.6375,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5632683664560318,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.6062753796577454,
+ "learning_rate": 0.00040141345970680814,
+ "loss": 0.5272740840911865,
+ "mean_token_accuracy": 0.8411559104919434,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5785442567273472,
+ "eval_loss": 0.6207635998725891,
+ "eval_mean_token_accuracy": 0.8259770478425401,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 100.5161,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5591374494135379,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.5771989226341248,
+ "learning_rate": 0.00039951233926444437,
+ "loss": 0.5237057685852051,
+ "mean_token_accuracy": 0.8413331478834152,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5765141543377651,
+ "eval_loss": 0.6119762063026428,
+ "eval_mean_token_accuracy": 0.8289416998289945,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 100.5288,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5470930114388466,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.4211709499359131,
+ "learning_rate": 0.00039755933813802937,
+ "loss": 0.5157970428466797,
+ "mean_token_accuracy": 0.8436284482479095,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5593592817528864,
+ "eval_loss": 0.6080028414726257,
+ "eval_mean_token_accuracy": 0.8277805718813049,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 100.495,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5636695981025696,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.6020221710205078,
+ "learning_rate": 0.00039555500900377994,
+ "loss": 0.5277415752410889,
+ "mean_token_accuracy": 0.840576496720314,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5837533748551701,
+ "eval_loss": 0.6048349142074585,
+ "eval_mean_token_accuracy": 0.8288552148288555,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5649972304701805,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.42798763513565063,
+ "learning_rate": 0.00039349991906313236,
+ "loss": 0.5282054901123047,
+ "mean_token_accuracy": 0.8409290999174118,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5473378684078709,
+ "eval_loss": 0.6050286889076233,
+ "eval_mean_token_accuracy": 0.8313557287950194,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 100.5342,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1140
+ },
+ {
+ "entropy": 0.5496659144759178,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.45835304260253906,
+ "learning_rate": 0.0003913946498822303,
+ "loss": 0.5158253669738769,
+ "mean_token_accuracy": 0.8436619535088539,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5719114836682094,
+ "eval_loss": 0.6046996116638184,
+ "eval_mean_token_accuracy": 0.8296863851922281,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 100.6617,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 1160
+ },
+ {
+ "entropy": 0.5694924771785737,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.5486838221549988,
+ "learning_rate": 0.000389239797227349,
+ "loss": 0.533592939376831,
+ "mean_token_accuracy": 0.8382544785737991,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5595886985237679,
+ "eval_loss": 0.6131696105003357,
+ "eval_mean_token_accuracy": 0.8292235069060594,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 100.5375,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5496176443994045,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.4574951231479645,
+ "learning_rate": 0.00038703597089630046,
+ "loss": 0.5157518863677979,
+ "mean_token_accuracy": 0.8423176407814026,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.57620483921485,
+ "eval_loss": 0.6032158732414246,
+ "eval_mean_token_accuracy": 0.8309619490350231,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 100.5207,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5365605391561985,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.6439656615257263,
+ "learning_rate": 0.00038478379454586755,
+ "loss": 0.5122960567474365,
+ "mean_token_accuracy": 0.8448946103453636,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5814327346474937,
+ "eval_loss": 0.6059017777442932,
+ "eval_mean_token_accuracy": 0.8299476896109206,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 100.6747,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5705336607992649,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.37888914346694946,
+ "learning_rate": 0.0003824839055153168,
+ "loss": 0.5322028160095215,
+ "mean_token_accuracy": 0.839797542989254,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5650541628344675,
+ "eval_loss": 0.5996877551078796,
+ "eval_mean_token_accuracy": 0.8318481569209796,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 100.4866,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1240
+ },
+ {
+ "entropy": 0.47180789709091187,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.48629021644592285,
+ "learning_rate": 0.00038013695464603875,
+ "loss": 0.4398338317871094,
+ "mean_token_accuracy": 0.8641970310455713,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5053895072320874,
+ "eval_loss": 0.6186225414276123,
+ "eval_mean_token_accuracy": 0.8313326363483172,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 100.6463,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1260
+ },
+ {
+ "entropy": 0.42782740518450735,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.6376653909683228,
+ "learning_rate": 0.00037774360609736765,
+ "loss": 0.39916558265686036,
+ "mean_token_accuracy": 0.87214947193861,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.4964009498277407,
+ "eval_loss": 0.6355555057525635,
+ "eval_mean_token_accuracy": 0.8277861790710621,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 100.5204,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1280
+ },
+ {
+ "entropy": 0.44719084799289704,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.44040077924728394,
+ "learning_rate": 0.00037530453715863216,
+ "loss": 0.41631083488464354,
+ "mean_token_accuracy": 0.8656514942646026,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.4866159513090434,
+ "eval_loss": 0.6143950819969177,
+ "eval_mean_token_accuracy": 0.8322170008434339,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 100.6652,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 1300
+ },
+ {
+ "entropy": 0.44159098640084266,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.5013471841812134,
+ "learning_rate": 0.00037282043805749067,
+ "loss": 0.41068315505981445,
+ "mean_token_accuracy": 0.8686433017253876,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5019330122832502,
+ "eval_loss": 0.6221247315406799,
+ "eval_mean_token_accuracy": 0.831573147787137,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 100.5258,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1320
+ },
+ {
+ "entropy": 0.4485637702047825,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5944355726242065,
+ "learning_rate": 0.0003702920117646045,
+ "loss": 0.4109023571014404,
+ "mean_token_accuracy": 0.8689506188035011,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.4770783360084791,
+ "eval_loss": 0.6268280148506165,
+ "eval_mean_token_accuracy": 0.8333997703000401,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 100.6448,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4550867825746536,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.564658522605896,
+ "learning_rate": 0.00036771997379470646,
+ "loss": 0.42194232940673826,
+ "mean_token_accuracy": 0.8678073868155479,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.49525917982787226,
+ "eval_loss": 0.6182692646980286,
+ "eval_mean_token_accuracy": 0.8333475599797924,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 100.5126,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4714259386062622,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.4722047746181488,
+ "learning_rate": 0.0003651050520041175,
+ "loss": 0.4292605400085449,
+ "mean_token_accuracy": 0.8639682114124299,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.5082498301615875,
+ "eval_loss": 0.6186839938163757,
+ "eval_mean_token_accuracy": 0.8319074859779872,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 100.4809,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1380
+ },
+ {
+ "entropy": 0.45790217220783236,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6115509271621704,
+ "learning_rate": 0.00036244798638477297,
+ "loss": 0.42031397819519045,
+ "mean_token_accuracy": 0.8663031980395317,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.48481654049305434,
+ "eval_loss": 0.6264435052871704,
+ "eval_mean_token_accuracy": 0.8322370792372843,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 100.4922,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 1400
+ },
+ {
+ "entropy": 0.45805501863360404,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.4944291412830353,
+ "learning_rate": 0.00035974952885481294,
+ "loss": 0.420308256149292,
+ "mean_token_accuracy": 0.8675160601735115,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.49597560086946807,
+ "eval_loss": 0.6137084364891052,
+ "eval_mean_token_accuracy": 0.8334959819075767,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 100.63,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4517170637845993,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4654815196990967,
+ "learning_rate": 0.0003570104430457988,
+ "loss": 0.4180137634277344,
+ "mean_token_accuracy": 0.8674369990825653,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.49280156814650206,
+ "eval_loss": 0.619944155216217,
+ "eval_mean_token_accuracy": 0.8338818181766553,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1440
+ },
+ {
+ "entropy": 0.45885001718997953,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6621031761169434,
+ "learning_rate": 0.00035423150408661424,
+ "loss": 0.4204216957092285,
+ "mean_token_accuracy": 0.8667945370078087,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5009692979327748,
+ "eval_loss": 0.6165568828582764,
+ "eval_mean_token_accuracy": 0.8340886737523454,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 100.5033,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1460
+ },
+ {
+ "entropy": 0.47144523113965986,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.6660031080245972,
+ "learning_rate": 0.000351413498384113,
+ "loss": 0.43781867027282717,
+ "mean_token_accuracy": 0.8613462626934052,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.49263788843422796,
+ "eval_loss": 0.6130704879760742,
+ "eval_mean_token_accuracy": 0.8341409915618683,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 100.475,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1480
+ },
+ {
+ "entropy": 0.4689751408994198,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.7863802909851074,
+ "learning_rate": 0.0003485572234005751,
+ "loss": 0.4308882713317871,
+ "mean_token_accuracy": 0.8638612285256386,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.46260853061515295,
+ "eval_loss": 0.6168972253799438,
+ "eval_mean_token_accuracy": 0.8346554998601421,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 100.4639,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1500
+ },
+ {
+ "entropy": 0.46667142286896707,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5704890489578247,
+ "learning_rate": 0.00034566348742803447,
+ "loss": 0.4309516429901123,
+ "mean_token_accuracy": 0.8622826755046844,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.4980781016389975,
+ "eval_loss": 0.6167429089546204,
+ "eval_mean_token_accuracy": 0.834536243355676,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 100.643,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 1520
+ },
+ {
+ "entropy": 0.46832413524389266,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.4302378296852112,
+ "learning_rate": 0.00034273310935954167,
+ "loss": 0.44164109230041504,
+ "mean_token_accuracy": 0.8621386975049973,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5203426120321403,
+ "eval_loss": 0.5976346135139465,
+ "eval_mean_token_accuracy": 0.835427654928036,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 100.6002,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1540
+ },
+ {
+ "entropy": 0.47617209553718565,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.3990112245082855,
+ "learning_rate": 0.000339766918457427,
+ "loss": 0.4339760303497314,
+ "mean_token_accuracy": 0.8615911722183227,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.5001180503475532,
+ "eval_loss": 0.5996196866035461,
+ "eval_mean_token_accuracy": 0.8364866047762753,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 100.563,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4649648956954479,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.561297595500946,
+ "learning_rate": 0.0003367657541186292,
+ "loss": 0.4299639701843262,
+ "mean_token_accuracy": 0.8626953899860382,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.4771459604916948,
+ "eval_loss": 0.6025190353393555,
+ "eval_mean_token_accuracy": 0.8365427228841888,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 100.4992,
+ "eval_samples_per_second": 14.129,
+ "eval_steps_per_second": 1.771,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4623337537050247,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5753264427185059,
+ "learning_rate": 0.0003337304656371558,
+ "loss": 0.4331684112548828,
+ "mean_token_accuracy": 0.8612382411956787,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.4919183887122722,
+ "eval_loss": 0.6000820398330688,
+ "eval_mean_token_accuracy": 0.8369202429658911,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 100.532,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1600
+ },
+ {
+ "entropy": 0.470283505320549,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.43974098563194275,
+ "learning_rate": 0.0003306619119637429,
+ "loss": 0.4338486671447754,
+ "mean_token_accuracy": 0.8619021520018577,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.49377609019199115,
+ "eval_loss": 0.5960330963134766,
+ "eval_mean_token_accuracy": 0.8362180245726296,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 100.4771,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1620
+ },
+ {
+ "entropy": 0.45613840222358704,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6422379612922668,
+ "learning_rate": 0.0003275609614627828,
+ "loss": 0.4256906032562256,
+ "mean_token_accuracy": 0.8665295109152794,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.498182592934437,
+ "eval_loss": 0.601266622543335,
+ "eval_mean_token_accuracy": 0.8354173841771115,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 100.5886,
+ "eval_samples_per_second": 14.117,
+ "eval_steps_per_second": 1.77,
+ "step": 1640
+ },
+ {
+ "entropy": 0.47032496562370885,
+ "epoch": 4.0,
+ "grad_norm": 1.9074139595031738,
+ "learning_rate": 0.0003244284916665865,
+ "loss": 0.4445479393005371,
+ "mean_token_accuracy": 0.8612772287466587,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.49686015171281406,
+ "eval_loss": 0.5976129770278931,
+ "eval_mean_token_accuracy": 0.8361410147018646,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 100.483,
+ "eval_samples_per_second": 14.132,
+ "eval_steps_per_second": 1.771,
+ "step": 1660
+ },
+ {
+ "entropy": 0.3599271237850189,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.3787924647331238,
+ "learning_rate": 0.0003212653890270527,
+ "loss": 0.31561665534973143,
+ "mean_token_accuracy": 0.895532414317131,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.42754596890358443,
+ "eval_loss": 0.6485731601715088,
+ "eval_mean_token_accuracy": 0.8336000315259012,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 100.6084,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 1680
+ },
+ {
+ "entropy": 0.3487696208059788,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.40927404165267944,
+ "learning_rate": 0.0003180725486648125,
+ "loss": 0.3065603494644165,
+ "mean_token_accuracy": 0.8975504830479621,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.39018024487441844,
+ "eval_loss": 0.6716603636741638,
+ "eval_mean_token_accuracy": 0.8346235541145454,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 100.6246,
+ "eval_samples_per_second": 14.112,
+ "eval_steps_per_second": 1.769,
+ "step": 1700
+ },
+ {
+ "entropy": 0.338431554287672,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.4601036012172699,
+ "learning_rate": 0.0003148508741159208,
+ "loss": 0.30837604999542234,
+ "mean_token_accuracy": 0.8966476947069169,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4045041859819648,
+ "eval_loss": 0.6581968665122986,
+ "eval_mean_token_accuracy": 0.8354246207837308,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 100.4883,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 1720
+ },
+ {
+ "entropy": 0.35083762779831884,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5361945033073425,
+ "learning_rate": 0.00031160127707616616,
+ "loss": 0.31211466789245607,
+ "mean_token_accuracy": 0.8945121049880982,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.41227720192309175,
+ "eval_loss": 0.6623457670211792,
+ "eval_mean_token_accuracy": 0.8348750354868643,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 100.6294,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 1740
+ },
+ {
+ "entropy": 0.3526027873158455,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.4194208085536957,
+ "learning_rate": 0.00030832467714307193,
+ "loss": 0.31311697959899903,
+ "mean_token_accuracy": 0.895439763367176,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.3938915096641926,
+ "eval_loss": 0.6631514430046082,
+ "eval_mean_token_accuracy": 0.8352231530660994,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 100.4578,
+ "eval_samples_per_second": 14.135,
+ "eval_steps_per_second": 1.772,
+ "step": 1760
+ },
+ {
+ "entropy": 0.3526589572429657,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.5007084012031555,
+ "learning_rate": 0.00030502200155566056,
+ "loss": 0.3219191551208496,
+ "mean_token_accuracy": 0.8929495573043823,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.4198530572518874,
+ "eval_loss": 0.6491358280181885,
+ "eval_mean_token_accuracy": 0.8345440529035718,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 100.4708,
+ "eval_samples_per_second": 14.133,
+ "eval_steps_per_second": 1.772,
+ "step": 1780
+ },
+ {
+ "entropy": 0.3524815313518047,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.49561530351638794,
+ "learning_rate": 0.00030169418493205646,
+ "loss": 0.31483731269836424,
+ "mean_token_accuracy": 0.8936498373746872,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.4056778658976716,
+ "eval_loss": 0.6486037373542786,
+ "eval_mean_token_accuracy": 0.8367906838990329,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 100.5668,
+ "eval_samples_per_second": 14.12,
+ "eval_steps_per_second": 1.77,
+ "step": 1800
+ },
+ {
+ "entropy": 0.3605773441493511,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.5805456638336182,
+ "learning_rate": 0.0002983421690049998,
+ "loss": 0.327486515045166,
+ "mean_token_accuracy": 0.8891258731484413,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.4375965514879548,
+ "eval_loss": 0.6490084528923035,
+ "eval_mean_token_accuracy": 0.8338087884227882,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 100.605,
+ "eval_samples_per_second": 14.115,
+ "eval_steps_per_second": 1.769,
+ "step": 1820
+ },
+ {
+ "entropy": 0.36023239344358443,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.4374412000179291,
+ "learning_rate": 0.00029496690235534753,
+ "loss": 0.32308504581451414,
+ "mean_token_accuracy": 0.8926876381039619,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.4166294440124812,
+ "eval_loss": 0.6568848490715027,
+ "eval_mean_token_accuracy": 0.8356464732898755,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 100.5314,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 1840
+ },
+ {
+ "entropy": 0.36440589353442193,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5468290448188782,
+ "learning_rate": 0.0002915693401436358,
+ "loss": 0.33129727840423584,
+ "mean_token_accuracy": 0.8899301409721374,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.4183989781677053,
+ "eval_loss": 0.6491609811782837,
+ "eval_mean_token_accuracy": 0.8347473757320576,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 100.5165,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 1860
+ },
+ {
+ "entropy": 0.35494092106819153,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5005297660827637,
+ "learning_rate": 0.0002881504438397814,
+ "loss": 0.3212761402130127,
+ "mean_token_accuracy": 0.8922176659107208,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.417422948593504,
+ "eval_loss": 0.6437649130821228,
+ "eval_mean_token_accuracy": 0.8363133615322327,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 100.4404,
+ "eval_samples_per_second": 14.138,
+ "eval_steps_per_second": 1.772,
+ "step": 1880
+ },
+ {
+ "entropy": 0.38629950806498525,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.5029445290565491,
+ "learning_rate": 0.00028471118095099653,
+ "loss": 0.3421678066253662,
+ "mean_token_accuracy": 0.8871301665902138,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.43428030978427845,
+ "eval_loss": 0.6387779712677002,
+ "eval_mean_token_accuracy": 0.8362574828474709,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 100.5615,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1900
+ },
+ {
+ "entropy": 0.35401932820677756,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.4358181357383728,
+ "learning_rate": 0.00028125252474799615,
+ "loss": 0.31899123191833495,
+ "mean_token_accuracy": 0.8915431290864945,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.44020974987678313,
+ "eval_loss": 0.6173001527786255,
+ "eval_mean_token_accuracy": 0.8383562400769652,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 100.5611,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 1920
+ },
+ {
+ "entropy": 0.36626047156751157,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6409969925880432,
+ "learning_rate": 0.00027777545398957387,
+ "loss": 0.3333898067474365,
+ "mean_token_accuracy": 0.8905606463551521,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4399699464942632,
+ "eval_loss": 0.6296016573905945,
+ "eval_mean_token_accuracy": 0.8374426036738278,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 100.525,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 1940
+ },
+ {
+ "entropy": 0.3736307479441166,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.584947943687439,
+ "learning_rate": 0.00027428095264562445,
+ "loss": 0.3333252191543579,
+ "mean_token_accuracy": 0.8869207128882408,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.42440525433990395,
+ "eval_loss": 0.6371653079986572,
+ "eval_mean_token_accuracy": 0.8369856202200557,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 100.4672,
+ "eval_samples_per_second": 14.134,
+ "eval_steps_per_second": 1.772,
+ "step": 1960
+ },
+ {
+ "entropy": 0.3753699027001858,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.6957218647003174,
+ "learning_rate": 0.0002707700096186923,
+ "loss": 0.3297238826751709,
+ "mean_token_accuracy": 0.8905770689249038,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.41796197449223377,
+ "eval_loss": 0.6490530371665955,
+ "eval_mean_token_accuracy": 0.8366815454504463,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 100.4524,
+ "eval_samples_per_second": 14.136,
+ "eval_steps_per_second": 1.772,
+ "step": 1980
+ },
+ {
+ "entropy": 0.3812362939119339,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.49150440096855164,
+ "learning_rate": 0.00026724361846412357,
+ "loss": 0.33749845027923586,
+ "mean_token_accuracy": 0.8862938195466995,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4048152939322289,
+ "eval_loss": 0.6516959071159363,
+ "eval_mean_token_accuracy": 0.8367159433579177,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 100.4853,
+ "eval_samples_per_second": 14.131,
+ "eval_steps_per_second": 1.771,
+ "step": 2000
+ },
+ {
+ "entropy": 0.36848620921373365,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.408181756734848,
+ "learning_rate": 0.0002637027771089018,
+ "loss": 0.33106796741485595,
+ "mean_token_accuracy": 0.8890365317463875,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4192614625678973,
+ "eval_loss": 0.6399418115615845,
+ "eval_mean_token_accuracy": 0.8374699645497826,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 100.5087,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2020
+ },
+ {
+ "entropy": 0.3796596243977547,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.44735926389694214,
+ "learning_rate": 0.0002601484875692464,
+ "loss": 0.3412761926651001,
+ "mean_token_accuracy": 0.8861836612224578,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.42973598469509167,
+ "eval_loss": 0.6350983381271362,
+ "eval_mean_token_accuracy": 0.8384529630119881,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 100.6357,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2040
+ },
+ {
+ "entropy": 0.37802436500787734,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.5981740355491638,
+ "learning_rate": 0.00025658175566705397,
+ "loss": 0.3330779790878296,
+ "mean_token_accuracy": 0.889920388162136,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4339508225073975,
+ "eval_loss": 0.6320109367370605,
+ "eval_mean_token_accuracy": 0.8392749485674869,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 100.6099,
+ "eval_samples_per_second": 14.114,
+ "eval_steps_per_second": 1.769,
+ "step": 2060
+ },
+ {
+ "entropy": 0.3353040623359191,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.37401750683784485,
+ "learning_rate": 0.0002530035907452624,
+ "loss": 0.28583431243896484,
+ "mean_token_accuracy": 0.9050691662690579,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.3525358152188612,
+ "eval_loss": 0.6934574842453003,
+ "eval_mean_token_accuracy": 0.8376689958438445,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 2080
+ },
+ {
+ "entropy": 0.2573637519031763,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.44206979870796204,
+ "learning_rate": 0.000249415005382219,
+ "loss": 0.21491749286651612,
+ "mean_token_accuracy": 0.9251434653997421,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.33512375880493206,
+ "eval_loss": 0.7306715250015259,
+ "eval_mean_token_accuracy": 0.8359865636638041,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 100.4978,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2100
+ },
+ {
+ "entropy": 0.2527348782867193,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.5964394807815552,
+ "learning_rate": 0.00024581701510513254,
+ "loss": 0.2123027563095093,
+ "mean_token_accuracy": 0.9265728950500488,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.33578530523214445,
+ "eval_loss": 0.7510417103767395,
+ "eval_mean_token_accuracy": 0.8321909639942512,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 100.658,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2120
+ },
+ {
+ "entropy": 0.24469284266233443,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.5237054228782654,
+ "learning_rate": 0.00024221063810269135,
+ "loss": 0.2136140823364258,
+ "mean_token_accuracy": 0.9270404040813446,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.35031337148687813,
+ "eval_loss": 0.7314468622207642,
+ "eval_mean_token_accuracy": 0.8337782072217277,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 100.6726,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2140
+ },
+ {
+ "entropy": 0.2677062630653381,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.42191237211227417,
+ "learning_rate": 0.0002385968949369278,
+ "loss": 0.22275702953338622,
+ "mean_token_accuracy": 0.924260924756527,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.34690580140338856,
+ "eval_loss": 0.7312859296798706,
+ "eval_mean_token_accuracy": 0.8356352070074403,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 100.5985,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2160
+ },
+ {
+ "entropy": 0.2557672385126352,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.5652918815612793,
+ "learning_rate": 0.00023497680825441142,
+ "loss": 0.21927871704101562,
+ "mean_token_accuracy": 0.9239626571536064,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.34665561459037697,
+ "eval_loss": 0.7144672870635986,
+ "eval_mean_token_accuracy": 0.8356724937310379,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 100.5245,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2180
+ },
+ {
+ "entropy": 0.25789538472890855,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.5560785531997681,
+ "learning_rate": 0.0002313514024968517,
+ "loss": 0.2210136890411377,
+ "mean_token_accuracy": 0.9243702575564384,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.35402616172024376,
+ "eval_loss": 0.7119109034538269,
+ "eval_mean_token_accuracy": 0.8353153925933196,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 100.5354,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.771,
+ "step": 2200
+ },
+ {
+ "entropy": 0.25272465534508226,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.5427626371383667,
+ "learning_rate": 0.00022772170361119376,
+ "loss": 0.2159494161605835,
+ "mean_token_accuracy": 0.9245263740420342,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.335209386067444,
+ "eval_loss": 0.7538248300552368,
+ "eval_mean_token_accuracy": 0.8342919965808311,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2220
+ },
+ {
+ "entropy": 0.2575274586677551,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.6143790483474731,
+ "learning_rate": 0.00022408873875928695,
+ "loss": 0.22290892601013185,
+ "mean_token_accuracy": 0.922516268491745,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.33513057499789123,
+ "eval_loss": 0.7243130803108215,
+ "eval_mean_token_accuracy": 0.8362914176469438,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 100.5302,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2240
+ },
+ {
+ "entropy": 0.2618753470480442,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.5207995176315308,
+ "learning_rate": 0.00022045353602721056,
+ "loss": 0.2217085361480713,
+ "mean_token_accuracy": 0.9243394300341606,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.34302107621444744,
+ "eval_loss": 0.7186807990074158,
+ "eval_mean_token_accuracy": 0.8363065294335398,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 100.5471,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2260
+ },
+ {
+ "entropy": 0.2617136687040329,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.6325564384460449,
+ "learning_rate": 0.00021681712413433747,
+ "loss": 0.22060952186584473,
+ "mean_token_accuracy": 0.9226485282182694,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.333045284064968,
+ "eval_loss": 0.7474223971366882,
+ "eval_mean_token_accuracy": 0.8359048500489653,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 100.5084,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2280
+ },
+ {
+ "entropy": 0.25311486050486565,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.5975027084350586,
+ "learning_rate": 0.00021318053214221873,
+ "loss": 0.21123769283294677,
+ "mean_token_accuracy": 0.9265317857265473,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.3362796777084972,
+ "eval_loss": 0.7327709197998047,
+ "eval_mean_token_accuracy": 0.837151965398467,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 100.5568,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2300
+ },
+ {
+ "entropy": 0.2633973427116871,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.50340336561203,
+ "learning_rate": 0.0002095447891633714,
+ "loss": 0.2288609504699707,
+ "mean_token_accuracy": 0.9207504242658615,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.34324210764986746,
+ "eval_loss": 0.720504641532898,
+ "eval_mean_token_accuracy": 0.8365555395571034,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 100.6339,
+ "eval_samples_per_second": 14.111,
+ "eval_steps_per_second": 1.769,
+ "step": 2320
+ },
+ {
+ "entropy": 0.2540372222661972,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.3917844593524933,
+ "learning_rate": 0.0002059109240700518,
+ "loss": 0.21576733589172364,
+ "mean_token_accuracy": 0.9244590282440186,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.32215716731682253,
+ "eval_loss": 0.7447317242622375,
+ "eval_mean_token_accuracy": 0.8370190708155043,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 100.5592,
+ "eval_samples_per_second": 14.121,
+ "eval_steps_per_second": 1.77,
+ "step": 2340
+ },
+ {
+ "entropy": 0.2568057753145695,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.5266931653022766,
+ "learning_rate": 0.0002022799652030969,
+ "loss": 0.22330741882324218,
+ "mean_token_accuracy": 0.9227217420935631,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.3442636997847075,
+ "eval_loss": 0.732654869556427,
+ "eval_mean_token_accuracy": 0.8347789558132043,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 100.5265,
+ "eval_samples_per_second": 14.126,
+ "eval_steps_per_second": 1.771,
+ "step": 2360
+ },
+ {
+ "entropy": 0.2599896214902401,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.5266039371490479,
+ "learning_rate": 0.0001986529400809158,
+ "loss": 0.22273578643798828,
+ "mean_token_accuracy": 0.923179779946804,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.315941037672959,
+ "eval_loss": 0.745282769203186,
+ "eval_mean_token_accuracy": 0.8377744347861643,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 100.6775,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2380
+ },
+ {
+ "entropy": 0.2554823886603117,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.4775709807872772,
+ "learning_rate": 0.00019503087510871467,
+ "loss": 0.21876795291900636,
+ "mean_token_accuracy": 0.9234203144907951,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.32689186612541754,
+ "eval_loss": 0.7430388927459717,
+ "eval_mean_token_accuracy": 0.8360160689675407,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 100.5173,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2400
+ },
+ {
+ "entropy": 0.2513945747166872,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.4408344626426697,
+ "learning_rate": 0.00019141479528803556,
+ "loss": 0.22275280952453613,
+ "mean_token_accuracy": 0.9230920314788819,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.3305554250987728,
+ "eval_loss": 0.719525933265686,
+ "eval_mean_token_accuracy": 0.8389456315656726,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 100.6401,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2420
+ },
+ {
+ "entropy": 0.25856637842953206,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.5113565921783447,
+ "learning_rate": 0.00018780572392669388,
+ "loss": 0.2229398727416992,
+ "mean_token_accuracy": 0.9219533413648605,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.3359096589383115,
+ "eval_loss": 0.7252099514007568,
+ "eval_mean_token_accuracy": 0.8376730104510703,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 100.5082,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2440
+ },
+ {
+ "entropy": 0.2600935664027929,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.5442810654640198,
+ "learning_rate": 0.00018420468234919352,
+ "loss": 0.2232747793197632,
+ "mean_token_accuracy": 0.9228743195533753,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.3390730025058382,
+ "eval_loss": 0.7105191349983215,
+ "eval_mean_token_accuracy": 0.8373022832897272,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 100.5781,
+ "eval_samples_per_second": 14.118,
+ "eval_steps_per_second": 1.77,
+ "step": 2460
+ },
+ {
+ "entropy": 0.26224863193929193,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.48302575945854187,
+ "learning_rate": 0.0001806126896077046,
+ "loss": 0.22573623657226563,
+ "mean_token_accuracy": 0.9222784146666527,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.3409589269522871,
+ "eval_loss": 0.7065299153327942,
+ "eval_mean_token_accuracy": 0.8383687798226818,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 100.5703,
+ "eval_samples_per_second": 14.119,
+ "eval_steps_per_second": 1.77,
+ "step": 2480
+ },
+ {
+ "entropy": 0.22054089720432574,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.4459492862224579,
+ "learning_rate": 0.000177030762193684,
+ "loss": 0.17556196451187134,
+ "mean_token_accuracy": 0.9398620357880225,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.2574445717455296,
+ "eval_loss": 0.8623653650283813,
+ "eval_mean_token_accuracy": 0.8354085784279899,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 100.5966,
+ "eval_samples_per_second": 14.116,
+ "eval_steps_per_second": 1.769,
+ "step": 2500
+ },
+ {
+ "entropy": 0.15792878065258265,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.36572420597076416,
+ "learning_rate": 0.0001734599137502196,
+ "loss": 0.1272625207901001,
+ "mean_token_accuracy": 0.9563913255929947,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.2690922211060363,
+ "eval_loss": 0.8302472829818726,
+ "eval_mean_token_accuracy": 0.8356359852163979,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 100.4963,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2520
+ },
+ {
+ "entropy": 0.16345136426389217,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.5506694316864014,
+ "learning_rate": 0.00016990115478518198,
+ "loss": 0.12678930759429932,
+ "mean_token_accuracy": 0.956980648636818,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.2540055900142434,
+ "eval_loss": 0.8733987212181091,
+ "eval_mean_token_accuracy": 0.8353191654333908,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 100.5416,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2540
+ },
+ {
+ "entropy": 0.16925445348024368,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.45296669006347656,
+ "learning_rate": 0.00016635549238526212,
+ "loss": 0.13375500440597535,
+ "mean_token_accuracy": 0.9541815564036369,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.26645253858204637,
+ "eval_loss": 0.8481063842773438,
+ "eval_mean_token_accuracy": 0.833757481213366,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 100.5309,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2560
+ },
+ {
+ "entropy": 0.16540961395949125,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.37874168157577515,
+ "learning_rate": 0.00016282392993097793,
+ "loss": 0.12654224634170533,
+ "mean_token_accuracy": 0.9559261187911033,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.26255769014693375,
+ "eval_loss": 0.8726609945297241,
+ "eval_mean_token_accuracy": 0.834410011098626,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 100.5326,
+ "eval_samples_per_second": 14.125,
+ "eval_steps_per_second": 1.771,
+ "step": 2580
+ },
+ {
+ "entropy": 0.16574543640017508,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.5216332077980042,
+ "learning_rate": 0.00015930746681272983,
+ "loss": 0.1316835641860962,
+ "mean_token_accuracy": 0.9549794599413872,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.25787094997221166,
+ "eval_loss": 0.8650625944137573,
+ "eval_mean_token_accuracy": 0.8351506855380669,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 100.6899,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 2600
+ },
+ {
+ "entropy": 0.16450108364224433,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.47154104709625244,
+ "learning_rate": 0.00015580709814798434,
+ "loss": 0.13114912509918214,
+ "mean_token_accuracy": 0.9550666123628616,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.26544143142325155,
+ "eval_loss": 0.8493700623512268,
+ "eval_mean_token_accuracy": 0.8354951072944684,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 100.537,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2620
+ },
+ {
+ "entropy": 0.15516163408756256,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.5267629027366638,
+ "learning_rate": 0.00015232381449966837,
+ "loss": 0.12384135723114013,
+ "mean_token_accuracy": 0.9578391805291175,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.2626389741395297,
+ "eval_loss": 0.8506026268005371,
+ "eval_mean_token_accuracy": 0.8361568866151102,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 100.5433,
+ "eval_samples_per_second": 14.123,
+ "eval_steps_per_second": 1.77,
+ "step": 2640
+ },
+ {
+ "entropy": 0.16577404085546732,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.5251840949058533,
+ "learning_rate": 0.00014885860159585105,
+ "loss": 0.13078413009643555,
+ "mean_token_accuracy": 0.9554133445024491,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.25928718446010957,
+ "eval_loss": 0.8621042370796204,
+ "eval_mean_token_accuracy": 0.8348304505428571,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 100.6374,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2660
+ },
+ {
+ "entropy": 0.17561652958393098,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.39062443375587463,
+ "learning_rate": 0.0001454124400507946,
+ "loss": 0.1369921326637268,
+ "mean_token_accuracy": 0.9532063737511635,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.2644507312372829,
+ "eval_loss": 0.8432015180587769,
+ "eval_mean_token_accuracy": 0.8353369570849987,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 100.6628,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 2680
+ },
+ {
+ "entropy": 0.1702792163938284,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.5517253875732422,
+ "learning_rate": 0.0001419863050874522,
+ "loss": 0.1334153175354004,
+ "mean_token_accuracy": 0.9544750079512596,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.2629738619990563,
+ "eval_loss": 0.851611316204071,
+ "eval_mean_token_accuracy": 0.8346863824330019,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 100.4976,
+ "eval_samples_per_second": 14.13,
+ "eval_steps_per_second": 1.771,
+ "step": 2700
+ },
+ {
+ "entropy": 0.1671909661963582,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.44552651047706604,
+ "learning_rate": 0.00013858116626149107,
+ "loss": 0.13589805364608765,
+ "mean_token_accuracy": 0.9533182814717293,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.2605011472373866,
+ "eval_loss": 0.8493867516517639,
+ "eval_mean_token_accuracy": 0.8359608690390427,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 100.6709,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 2720
+ },
+ {
+ "entropy": 0.1699025196954608,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.5227566361427307,
+ "learning_rate": 0.0001351979871869202,
+ "loss": 0.1365843653678894,
+ "mean_token_accuracy": 0.9541483774781228,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.2580871745274308,
+ "eval_loss": 0.8643819689750671,
+ "eval_mean_token_accuracy": 0.8363788667019834,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 100.5146,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2740
+ },
+ {
+ "entropy": 0.17232896946370602,
+ "epoch": 6.651387213510254,
+ "grad_norm": 0.5396640300750732,
+ "learning_rate": 0.0001318377252633985,
+ "loss": 0.13684563636779784,
+ "mean_token_accuracy": 0.9534787341952324,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.26042155360572794,
+ "eval_loss": 0.8471524119377136,
+ "eval_mean_token_accuracy": 0.8368232735757077,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 100.6436,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2760
+ },
+ {
+ "entropy": 0.16852661557495593,
+ "epoch": 6.699638118214716,
+ "grad_norm": 0.5081996917724609,
+ "learning_rate": 0.00012850133140530215,
+ "loss": 0.13659658432006835,
+ "mean_token_accuracy": 0.9539226442575455,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.2588982036274471,
+ "eval_loss": 0.8474684357643127,
+ "eval_mean_token_accuracy": 0.8362404295567716,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 100.5156,
+ "eval_samples_per_second": 14.127,
+ "eval_steps_per_second": 1.771,
+ "step": 2780
+ },
+ {
+ "entropy": 0.17256609313189983,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.42446374893188477,
+ "learning_rate": 0.00012518974977262744,
+ "loss": 0.13608640432357788,
+ "mean_token_accuracy": 0.9528463885188103,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.2609142000588138,
+ "eval_loss": 0.8551791310310364,
+ "eval_mean_token_accuracy": 0.8362516620185938,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 100.5105,
+ "eval_samples_per_second": 14.128,
+ "eval_steps_per_second": 1.771,
+ "step": 2800
+ },
+ {
+ "entropy": 0.16441288813948632,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.4241557717323303,
+ "learning_rate": 0.00012190391750380397,
+ "loss": 0.13090516328811647,
+ "mean_token_accuracy": 0.9552061229944229,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.26902279090345577,
+ "eval_loss": 0.847598671913147,
+ "eval_mean_token_accuracy": 0.8359841473316878,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 100.6468,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2820
+ },
+ {
+ "entropy": 0.163655917160213,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.5481321215629578,
+ "learning_rate": 0.00011864476445049639,
+ "loss": 0.1307414174079895,
+ "mean_token_accuracy": 0.9541522398591041,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.26397091167026693,
+ "eval_loss": 0.8470706343650818,
+ "eval_mean_token_accuracy": 0.8352796468841895,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 100.5411,
+ "eval_samples_per_second": 14.124,
+ "eval_steps_per_second": 1.77,
+ "step": 2840
+ },
+ {
+ "entropy": 0.17050919625908137,
+ "epoch": 6.892641737032569,
+ "grad_norm": 0.6102217435836792,
+ "learning_rate": 0.00011541321291446658,
+ "loss": 0.13289833068847656,
+ "mean_token_accuracy": 0.9534440621733665,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.2447421256076084,
+ "eval_loss": 0.8822773098945618,
+ "eval_mean_token_accuracy": 0.8370943755915995,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 100.6409,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 2860
+ },
+ {
+ "entropy": 0.16003708317875862,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.4826158285140991,
+ "learning_rate": 0.00011221017738657364,
+ "loss": 0.13135261535644532,
+ "mean_token_accuracy": 0.9555956274271011,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.25742107462347225,
+ "eval_loss": 0.8532347679138184,
+ "eval_mean_token_accuracy": 0.836660507689701,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 100.6435,
+ "eval_samples_per_second": 14.109,
+ "eval_steps_per_second": 1.769,
+ "step": 2880
+ },
+ {
+ "entropy": 0.16182902473956345,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.4449099600315094,
+ "learning_rate": 0.00010903656428798382,
+ "loss": 0.1279471516609192,
+ "mean_token_accuracy": 0.9556913211941719,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.2602201160252764,
+ "eval_loss": 0.8557486534118652,
+ "eval_mean_token_accuracy": 0.8361740815505553,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 100.7992,
+ "eval_samples_per_second": 14.087,
+ "eval_steps_per_second": 1.766,
+ "step": 2900
+ },
+ {
+ "entropy": 0.1400035818417867,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.36720654368400574,
+ "learning_rate": 0.00010589327171366368,
+ "loss": 0.09224887490272522,
+ "mean_token_accuracy": 0.9685117831596961,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.21291789864556174,
+ "eval_loss": 0.9790643453598022,
+ "eval_mean_token_accuracy": 0.8361084946755613,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 100.7775,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 2920
+ },
+ {
+ "entropy": 0.10825499575585126,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.2915419936180115,
+ "learning_rate": 0.00010278118917823028,
+ "loss": 0.07976844906806946,
+ "mean_token_accuracy": 0.9731380000710488,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.21083223367674966,
+ "eval_loss": 0.9892836213111877,
+ "eval_mean_token_accuracy": 0.836239791318272,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 100.9786,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 2940
+ },
+ {
+ "entropy": 0.11260611563920975,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.3968478739261627,
+ "learning_rate": 9.970119736422877e-05,
+ "loss": 0.08049036264419555,
+ "mean_token_accuracy": 0.9727092757821083,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.21443948710567495,
+ "eval_loss": 0.9841322302818298,
+ "eval_mean_token_accuracy": 0.8369794662079114,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 2960
+ },
+ {
+ "entropy": 0.115106151252985,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.30111560225486755,
+ "learning_rate": 9.665416787290954e-05,
+ "loss": 0.08258993625640869,
+ "mean_token_accuracy": 0.9720528662204743,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.20990465422359744,
+ "eval_loss": 0.9941005706787109,
+ "eval_mean_token_accuracy": 0.8368785672643212,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 100.8138,
+ "eval_samples_per_second": 14.085,
+ "eval_steps_per_second": 1.766,
+ "step": 2980
+ },
+ {
+ "entropy": 0.10764018073678017,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.24775347113609314,
+ "learning_rate": 9.3640962977576e-05,
+ "loss": 0.07928993701934814,
+ "mean_token_accuracy": 0.9743952095508576,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.20923426203178555,
+ "eval_loss": 0.9848349094390869,
+ "eval_mean_token_accuracy": 0.8376054067290231,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 100.7195,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3000
+ },
+ {
+ "entropy": 0.11157451178878545,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.33588629961013794,
+ "learning_rate": 9.066243537957052e-05,
+ "loss": 0.08099400997161865,
+ "mean_token_accuracy": 0.9726107493042946,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.20457115518243124,
+ "eval_loss": 1.007703423500061,
+ "eval_mean_token_accuracy": 0.836432515570287,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 100.8888,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3020
+ },
+ {
+ "entropy": 0.10864771818742155,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.43094608187675476,
+ "learning_rate": 8.771942796697095e-05,
+ "loss": 0.0833050012588501,
+ "mean_token_accuracy": 0.9727370664477348,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.21487541170267577,
+ "eval_loss": 0.9794149994850159,
+ "eval_mean_token_accuracy": 0.8368617674607909,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 100.6794,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3040
+ },
+ {
+ "entropy": 0.11729530803859234,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.26597073674201965,
+ "learning_rate": 8.481277357606256e-05,
+ "loss": 0.0843969702720642,
+ "mean_token_accuracy": 0.9717461839318275,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.2150757516200623,
+ "eval_loss": 0.9721782803535461,
+ "eval_mean_token_accuracy": 0.8368598500664315,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 100.7485,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3060
+ },
+ {
+ "entropy": 0.11559880636632443,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.3401016294956207,
+ "learning_rate": 8.194329475565544e-05,
+ "loss": 0.08105930685997009,
+ "mean_token_accuracy": 0.9723709270358085,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.21134269748176082,
+ "eval_loss": 0.9862396717071533,
+ "eval_mean_token_accuracy": 0.8374697743506914,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 100.7679,
+ "eval_samples_per_second": 14.092,
+ "eval_steps_per_second": 1.766,
+ "step": 3080
+ },
+ {
+ "entropy": 0.11348765343427658,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.3345467150211334,
+ "learning_rate": 7.91118035343133e-05,
+ "loss": 0.08072147965431213,
+ "mean_token_accuracy": 0.9729523852467536,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.2079042781437381,
+ "eval_loss": 0.9928281307220459,
+ "eval_mean_token_accuracy": 0.8369294528880816,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 100.6961,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3100
+ },
+ {
+ "entropy": 0.10812166854739189,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.4052855968475342,
+ "learning_rate": 7.63191011905583e-05,
+ "loss": 0.07858278155326844,
+ "mean_token_accuracy": 0.9745768919587136,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.21274138660578246,
+ "eval_loss": 0.979730486869812,
+ "eval_mean_token_accuracy": 0.8359713946165663,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 100.8712,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3120
+ },
+ {
+ "entropy": 0.10486099142581225,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.3344109058380127,
+ "learning_rate": 7.35659780261194e-05,
+ "loss": 0.07731404900550842,
+ "mean_token_accuracy": 0.9740523621439934,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.21002410996830864,
+ "eval_loss": 0.9813475012779236,
+ "eval_mean_token_accuracy": 0.8367562504966607,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 100.6853,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3140
+ },
+ {
+ "entropy": 0.1151945998892188,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.35936611890792847,
+ "learning_rate": 7.085321314228569e-05,
+ "loss": 0.0848338782787323,
+ "mean_token_accuracy": 0.9722493141889572,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.21051086234242727,
+ "eval_loss": 0.9826087355613708,
+ "eval_mean_token_accuracy": 0.8371691442607494,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 100.7106,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3160
+ },
+ {
+ "entropy": 0.11743231695145369,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.2966472804546356,
+ "learning_rate": 6.818157421943049e-05,
+ "loss": 0.08336361646652221,
+ "mean_token_accuracy": 0.9715887486934662,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.2132127285003662,
+ "eval_loss": 0.9794583320617676,
+ "eval_mean_token_accuracy": 0.836922631839688,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 100.8284,
+ "eval_samples_per_second": 14.083,
+ "eval_steps_per_second": 1.765,
+ "step": 3180
+ },
+ {
+ "entropy": 0.12011048756539822,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.41335737705230713,
+ "learning_rate": 6.555181729976638e-05,
+ "loss": 0.08217214345932007,
+ "mean_token_accuracy": 0.9720462530851364,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.20867667121163916,
+ "eval_loss": 1.0004463195800781,
+ "eval_mean_token_accuracy": 0.8364931020174133,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 100.7152,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3200
+ },
+ {
+ "entropy": 0.11083896011114121,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.3560662865638733,
+ "learning_rate": 6.296468657339478e-05,
+ "loss": 0.07983347177505493,
+ "mean_token_accuracy": 0.9731158778071404,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.20952951665339845,
+ "eval_loss": 0.9946833848953247,
+ "eval_mean_token_accuracy": 0.8366418815730663,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 100.718,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3220
+ },
+ {
+ "entropy": 0.10758460871875286,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.23847290873527527,
+ "learning_rate": 6.0420914167708276e-05,
+ "loss": 0.07738233804702759,
+ "mean_token_accuracy": 0.973819300532341,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.20503406025720447,
+ "eval_loss": 1.0157110691070557,
+ "eval_mean_token_accuracy": 0.8368847350056252,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 100.712,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3240
+ },
+ {
+ "entropy": 0.11236482234671712,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.3121676445007324,
+ "learning_rate": 5.7921219940207895e-05,
+ "loss": 0.07914502620697021,
+ "mean_token_accuracy": 0.973124249279499,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.20229623801587673,
+ "eval_loss": 1.027987003326416,
+ "eval_mean_token_accuracy": 0.8369987684019496,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 100.6693,
+ "eval_samples_per_second": 14.106,
+ "eval_steps_per_second": 1.768,
+ "step": 3260
+ },
+ {
+ "entropy": 0.11417193301022052,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.3817048966884613,
+ "learning_rate": 5.546631127479136e-05,
+ "loss": 0.08328820466995239,
+ "mean_token_accuracy": 0.971915477514267,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.210720699406072,
+ "eval_loss": 1.0018188953399658,
+ "eval_mean_token_accuracy": 0.8371731913491581,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 100.6761,
+ "eval_samples_per_second": 14.105,
+ "eval_steps_per_second": 1.768,
+ "step": 3280
+ },
+ {
+ "entropy": 0.11353797391057015,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.441937655210495,
+ "learning_rate": 5.305688288157232e-05,
+ "loss": 0.08037771582603455,
+ "mean_token_accuracy": 0.973035278916359,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.20455422168702223,
+ "eval_loss": 1.012039065361023,
+ "eval_mean_token_accuracy": 0.837402118056008,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 100.755,
+ "eval_samples_per_second": 14.094,
+ "eval_steps_per_second": 1.767,
+ "step": 3300
+ },
+ {
+ "entropy": 0.10873433068776742,
+ "epoch": 8.0,
+ "grad_norm": 1.0863004922866821,
+ "learning_rate": 5.069361660028519e-05,
+ "loss": 0.0815244197845459,
+ "mean_token_accuracy": 0.9725936299715287,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.20847410825865992,
+ "eval_loss": 1.0063199996948242,
+ "eval_mean_token_accuracy": 0.8363274319118328,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 100.762,
+ "eval_samples_per_second": 14.093,
+ "eval_steps_per_second": 1.767,
+ "step": 3320
+ },
+ {
+ "entropy": 0.1041649229824543,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.31490007042884827,
+ "learning_rate": 4.8377181207332886e-05,
+ "loss": 0.06518369913101196,
+ "mean_token_accuracy": 0.9779868647456169,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.1915324301029859,
+ "eval_loss": 1.0649940967559814,
+ "eval_mean_token_accuracy": 0.8372375332237629,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 100.7428,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3340
+ },
+ {
+ "entropy": 0.09381815856322646,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.24123582243919373,
+ "learning_rate": 4.610823222653079e-05,
+ "loss": 0.06284829378128051,
+ "mean_token_accuracy": 0.9781818613409996,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.18765907483489325,
+ "eval_loss": 1.0903619527816772,
+ "eval_mean_token_accuracy": 0.8366766642318683,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 100.6893,
+ "eval_samples_per_second": 14.103,
+ "eval_steps_per_second": 1.768,
+ "step": 3360
+ },
+ {
+ "entropy": 0.0951899609528482,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.14863838255405426,
+ "learning_rate": 4.388741174360135e-05,
+ "loss": 0.06521615982055665,
+ "mean_token_accuracy": 0.9779476702213288,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.18881698044833173,
+ "eval_loss": 1.092907428741455,
+ "eval_mean_token_accuracy": 0.8373543248417672,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 100.8071,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3380
+ },
+ {
+ "entropy": 0.08654136713594199,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.21708497405052185,
+ "learning_rate": 4.171534822447155e-05,
+ "loss": 0.05802627205848694,
+ "mean_token_accuracy": 0.9801685467362404,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.1903750839528073,
+ "eval_loss": 1.0906604528427124,
+ "eval_mean_token_accuracy": 0.8376733345931835,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 100.6938,
+ "eval_samples_per_second": 14.102,
+ "eval_steps_per_second": 1.768,
+ "step": 3400
+ },
+ {
+ "entropy": 0.09255459355190396,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.1456836760044098,
+ "learning_rate": 3.959265633742418e-05,
+ "loss": 0.06009572744369507,
+ "mean_token_accuracy": 0.9788642093539238,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.18949040293358685,
+ "eval_loss": 1.0861941576004028,
+ "eval_mean_token_accuracy": 0.8378280066372303,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 100.6998,
+ "eval_samples_per_second": 14.101,
+ "eval_steps_per_second": 1.768,
+ "step": 3420
+ },
+ {
+ "entropy": 0.10507828444242477,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.14788983762264252,
+ "learning_rate": 3.751993677915433e-05,
+ "loss": 0.0709911584854126,
+ "mean_token_accuracy": 0.9762946248054505,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.185680764361044,
+ "eval_loss": 1.0962878465652466,
+ "eval_mean_token_accuracy": 0.8378534531325437,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 100.6613,
+ "eval_samples_per_second": 14.107,
+ "eval_steps_per_second": 1.768,
+ "step": 3440
+ },
+ {
+ "entropy": 0.0943055517040193,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.22308449447155,
+ "learning_rate": 3.549777610477903e-05,
+ "loss": 0.061539626121520995,
+ "mean_token_accuracy": 0.9780292823910713,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.1858175231164761,
+ "eval_loss": 1.102445363998413,
+ "eval_mean_token_accuracy": 0.8375556921021322,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 100.679,
+ "eval_samples_per_second": 14.104,
+ "eval_steps_per_second": 1.768,
+ "step": 3460
+ },
+ {
+ "entropy": 0.09218643279746175,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.30906444787979126,
+ "learning_rate": 3.352674656184925e-05,
+ "loss": 0.06346401572227478,
+ "mean_token_accuracy": 0.9781965032219887,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.1874545087640205,
+ "eval_loss": 1.1002845764160156,
+ "eval_mean_token_accuracy": 0.8373500513896514,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 100.6344,
+ "eval_samples_per_second": 14.11,
+ "eval_steps_per_second": 1.769,
+ "step": 3480
+ },
+ {
+ "entropy": 0.09471671236678958,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.17976395785808563,
+ "learning_rate": 3.1607405928410694e-05,
+ "loss": 0.0637880802154541,
+ "mean_token_accuracy": 0.9773588865995407,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.1869328013631735,
+ "eval_loss": 1.1063703298568726,
+ "eval_mean_token_accuracy": 0.8371504943692283,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 100.7793,
+ "eval_samples_per_second": 14.09,
+ "eval_steps_per_second": 1.766,
+ "step": 3500
+ },
+ {
+ "entropy": 0.09159188224002719,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.12867353856563568,
+ "learning_rate": 2.9740297355158894e-05,
+ "loss": 0.06207664012908935,
+ "mean_token_accuracy": 0.9785440415143967,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.18606380540668294,
+ "eval_loss": 1.1082826852798462,
+ "eval_mean_token_accuracy": 0.8376104299272045,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 100.6522,
+ "eval_samples_per_second": 14.108,
+ "eval_steps_per_second": 1.768,
+ "step": 3520
+ },
+ {
+ "entropy": 0.09101387113332748,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.26646703481674194,
+ "learning_rate": 2.792594921173444e-05,
+ "loss": 0.06317077279090881,
+ "mean_token_accuracy": 0.9785649806261063,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.18594747279467208,
+ "eval_loss": 1.1053133010864258,
+ "eval_mean_token_accuracy": 0.8377194324236238,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 100.7373,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3540
+ },
+ {
+ "entropy": 0.09201286779716611,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.16690310835838318,
+ "learning_rate": 2.6164874937200204e-05,
+ "loss": 0.06450713276863099,
+ "mean_token_accuracy": 0.9791284471750259,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.18545364337355902,
+ "eval_loss": 1.1091718673706055,
+ "eval_mean_token_accuracy": 0.8377593213252807,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 100.7371,
+ "eval_samples_per_second": 14.096,
+ "eval_steps_per_second": 1.767,
+ "step": 3560
+ },
+ {
+ "entropy": 0.09645880302414298,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.18617478013038635,
+ "learning_rate": 2.4457572894744302e-05,
+ "loss": 0.06452478170394897,
+ "mean_token_accuracy": 0.9772841408848763,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.18613224945376428,
+ "eval_loss": 1.112298607826233,
+ "eval_mean_token_accuracy": 0.8372736559824997,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 100.9834,
+ "eval_samples_per_second": 14.062,
+ "eval_steps_per_second": 1.763,
+ "step": 3580
+ },
+ {
+ "entropy": 0.09433056097477674,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.2585618793964386,
+ "learning_rate": 2.2804526230649412e-05,
+ "loss": 0.0643742561340332,
+ "mean_token_accuracy": 0.9784440502524376,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.18546074536744128,
+ "eval_loss": 1.1133522987365723,
+ "eval_mean_token_accuracy": 0.8377416592635466,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 101.0253,
+ "eval_samples_per_second": 14.056,
+ "eval_steps_per_second": 1.762,
+ "step": 3600
+ },
+ {
+ "entropy": 0.0894506118260324,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.19618846476078033,
+ "learning_rate": 2.120620273756776e-05,
+ "loss": 0.06242886781692505,
+ "mean_token_accuracy": 0.9787433609366417,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.18585375219248654,
+ "eval_loss": 1.1109825372695923,
+ "eval_mean_token_accuracy": 0.8377892870581551,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 100.974,
+ "eval_samples_per_second": 14.063,
+ "eval_steps_per_second": 1.763,
+ "step": 3620
+ },
+ {
+ "entropy": 0.09536345470696687,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.10027517378330231,
+ "learning_rate": 1.9663054722141587e-05,
+ "loss": 0.0623970091342926,
+ "mean_token_accuracy": 0.9781744226813316,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.18505671804540613,
+ "eval_loss": 1.1121324300765991,
+ "eval_mean_token_accuracy": 0.8378979500090138,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 100.9466,
+ "eval_samples_per_second": 14.067,
+ "eval_steps_per_second": 1.763,
+ "step": 3640
+ },
+ {
+ "entropy": 0.09590104129165411,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.19535337388515472,
+ "learning_rate": 1.8175518877005552e-05,
+ "loss": 0.06799691319465637,
+ "mean_token_accuracy": 0.9761223942041397,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.18593402583612484,
+ "eval_loss": 1.1094188690185547,
+ "eval_mean_token_accuracy": 0.8378441116113341,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 100.8798,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3660
+ },
+ {
+ "entropy": 0.0912214593961835,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.16693317890167236,
+ "learning_rate": 1.6744016157207948e-05,
+ "loss": 0.06410449743270874,
+ "mean_token_accuracy": 0.9774021029472351,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.18701685898089676,
+ "eval_loss": 1.1080079078674316,
+ "eval_mean_token_accuracy": 0.8379885400279184,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 100.8862,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3680
+ },
+ {
+ "entropy": 0.09622923573479056,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.15017707645893097,
+ "learning_rate": 1.53689516610855e-05,
+ "loss": 0.06397625803947449,
+ "mean_token_accuracy": 0.9777478039264679,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.1863192869873529,
+ "eval_loss": 1.1090794801712036,
+ "eval_mean_token_accuracy": 0.8382032892007506,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 100.8767,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3700
+ },
+ {
+ "entropy": 0.0912615241482854,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.1881953328847885,
+ "learning_rate": 1.4050714515625077e-05,
+ "loss": 0.062059921026229856,
+ "mean_token_accuracy": 0.9778110668063164,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.18606430442815416,
+ "eval_loss": 1.109899640083313,
+ "eval_mean_token_accuracy": 0.8379705333977603,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 100.7437,
+ "eval_samples_per_second": 14.095,
+ "eval_steps_per_second": 1.767,
+ "step": 3720
+ },
+ {
+ "entropy": 0.09705390800268222,
+ "epoch": 9.012062726176115,
+ "grad_norm": 0.3272952437400818,
+ "learning_rate": 1.2789677766345677e-05,
+ "loss": 0.06879411935806275,
+ "mean_token_accuracy": 0.9774823463880099,
+ "num_tokens": 9122597.0,
+ "step": 3740
+ },
+ {
+ "epoch": 9.012062726176115,
+ "eval_entropy": 0.18716304565078756,
+ "eval_loss": 1.103650689125061,
+ "eval_mean_token_accuracy": 0.8379069282097763,
+ "eval_num_tokens": 9122597.0,
+ "eval_runtime": 100.8591,
+ "eval_samples_per_second": 14.079,
+ "eval_steps_per_second": 1.765,
+ "step": 3740
+ },
+ {
+ "entropy": 0.09239197438582777,
+ "epoch": 9.060313630880579,
+ "grad_norm": 0.12848101556301117,
+ "learning_rate": 1.1586198271730714e-05,
+ "loss": 0.058521777391433716,
+ "mean_token_accuracy": 0.9808280184864998,
+ "num_tokens": 9171477.0,
+ "step": 3760
+ },
+ {
+ "epoch": 9.060313630880579,
+ "eval_entropy": 0.18443025656965342,
+ "eval_loss": 1.1128716468811035,
+ "eval_mean_token_accuracy": 0.8376701189560837,
+ "eval_num_tokens": 9171477.0,
+ "eval_runtime": 100.7759,
+ "eval_samples_per_second": 14.091,
+ "eval_steps_per_second": 1.766,
+ "step": 3760
+ },
+ {
+ "entropy": 0.09276993470266462,
+ "epoch": 9.108564535585042,
+ "grad_norm": 0.0987873300909996,
+ "learning_rate": 1.0440616602241303e-05,
+ "loss": 0.06252194046974183,
+ "mean_token_accuracy": 0.9791185334324837,
+ "num_tokens": 9217543.0,
+ "step": 3780
+ },
+ {
+ "epoch": 9.108564535585042,
+ "eval_entropy": 0.18163840397355263,
+ "eval_loss": 1.124428629875183,
+ "eval_mean_token_accuracy": 0.8379190537366974,
+ "eval_num_tokens": 9217543.0,
+ "eval_runtime": 100.7229,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3780
+ },
+ {
+ "entropy": 0.08816529829055071,
+ "epoch": 9.156815440289506,
+ "grad_norm": 0.10476374626159668,
+ "learning_rate": 9.353256943939065e-06,
+ "loss": 0.056501579284667966,
+ "mean_token_accuracy": 0.9800700470805168,
+ "num_tokens": 9268724.0,
+ "step": 3800
+ },
+ {
+ "epoch": 9.156815440289506,
+ "eval_entropy": 0.17974376218037658,
+ "eval_loss": 1.1325938701629639,
+ "eval_mean_token_accuracy": 0.8377991332097,
+ "eval_num_tokens": 9268724.0,
+ "eval_runtime": 100.8729,
+ "eval_samples_per_second": 14.077,
+ "eval_steps_per_second": 1.765,
+ "step": 3800
+ },
+ {
+ "entropy": 0.0886100891046226,
+ "epoch": 9.205066344993968,
+ "grad_norm": 0.10369760543107986,
+ "learning_rate": 8.32442700674502e-06,
+ "loss": 0.05948063135147095,
+ "mean_token_accuracy": 0.9798692613840103,
+ "num_tokens": 9316805.0,
+ "step": 3820
+ },
+ {
+ "epoch": 9.205066344993968,
+ "eval_entropy": 0.1793072780364015,
+ "eval_loss": 1.1363310813903809,
+ "eval_mean_token_accuracy": 0.8378076965219519,
+ "eval_num_tokens": 9316805.0,
+ "eval_runtime": 100.8909,
+ "eval_samples_per_second": 14.075,
+ "eval_steps_per_second": 1.764,
+ "step": 3820
+ },
+ {
+ "entropy": 0.07910045385360717,
+ "epoch": 9.253317249698432,
+ "grad_norm": 0.06480544060468674,
+ "learning_rate": 7.3544179373615505e-06,
+ "loss": 0.053755837678909305,
+ "mean_token_accuracy": 0.9819905251264572,
+ "num_tokens": 9369940.0,
+ "step": 3840
+ },
+ {
+ "epoch": 9.253317249698432,
+ "eval_entropy": 0.1782708638086078,
+ "eval_loss": 1.1404398679733276,
+ "eval_mean_token_accuracy": 0.8379109310969878,
+ "eval_num_tokens": 9369940.0,
+ "eval_runtime": 100.7264,
+ "eval_samples_per_second": 14.098,
+ "eval_steps_per_second": 1.767,
+ "step": 3840
+ },
+ {
+ "entropy": 0.09386134278029204,
+ "epoch": 9.301568154402895,
+ "grad_norm": 0.12832233309745789,
+ "learning_rate": 6.443504236881163e-06,
+ "loss": 0.061395853757858276,
+ "mean_token_accuracy": 0.9780228048563003,
+ "num_tokens": 9416788.0,
+ "step": 3860
+ },
+ {
+ "epoch": 9.301568154402895,
+ "eval_entropy": 0.1777799724276816,
+ "eval_loss": 1.1431900262832642,
+ "eval_mean_token_accuracy": 0.8377661607908399,
+ "eval_num_tokens": 9416788.0,
+ "eval_runtime": 100.8637,
+ "eval_samples_per_second": 14.078,
+ "eval_steps_per_second": 1.765,
+ "step": 3860
+ },
+ {
+ "entropy": 0.0827090598642826,
+ "epoch": 9.349819059107359,
+ "grad_norm": 0.12366542220115662,
+ "learning_rate": 5.591943683106039e-06,
+ "loss": 0.05561224818229675,
+ "mean_token_accuracy": 0.9803308844566345,
+ "num_tokens": 9467862.0,
+ "step": 3880
+ },
+ {
+ "epoch": 9.349819059107359,
+ "eval_entropy": 0.1778556579368168,
+ "eval_loss": 1.1441748142242432,
+ "eval_mean_token_accuracy": 0.8377985549107027,
+ "eval_num_tokens": 9467862.0,
+ "eval_runtime": 100.7182,
+ "eval_samples_per_second": 14.099,
+ "eval_steps_per_second": 1.767,
+ "step": 3880
+ },
+ {
+ "entropy": 0.09148634811863303,
+ "epoch": 9.398069963811821,
+ "grad_norm": 0.1590287685394287,
+ "learning_rate": 4.799977257599879e-06,
+ "loss": 0.06154900789260864,
+ "mean_token_accuracy": 0.9787917152047158,
+ "num_tokens": 9514611.0,
+ "step": 3900
+ },
+ {
+ "epoch": 9.398069963811821,
+ "eval_entropy": 0.17795597113082917,
+ "eval_loss": 1.1449599266052246,
+ "eval_mean_token_accuracy": 0.8378339922160245,
+ "eval_num_tokens": 9514611.0,
+ "eval_runtime": 100.8791,
+ "eval_samples_per_second": 14.076,
+ "eval_steps_per_second": 1.764,
+ "step": 3900
+ },
+ {
+ "entropy": 0.0879560124129057,
+ "epoch": 9.446320868516285,
+ "grad_norm": 0.11894718557596207,
+ "learning_rate": 4.067829077493076e-06,
+ "loss": 0.05804988145828247,
+ "mean_token_accuracy": 0.9796567305922508,
+ "num_tokens": 9563437.0,
+ "step": 3920
+ },
+ {
+ "epoch": 9.446320868516285,
+ "eval_entropy": 0.17797360205081073,
+ "eval_loss": 1.146341323852539,
+ "eval_mean_token_accuracy": 0.8379327249660921,
+ "eval_num_tokens": 9563437.0,
+ "eval_runtime": 100.8092,
+ "eval_samples_per_second": 14.086,
+ "eval_steps_per_second": 1.766,
+ "step": 3920
+ },
+ {
+ "entropy": 0.0873468822799623,
+ "epoch": 9.494571773220748,
+ "grad_norm": 0.09161137789487839,
+ "learning_rate": 3.3957063320601047e-06,
+ "loss": 0.05694714784622192,
+ "mean_token_accuracy": 0.9784631252288818,
+ "num_tokens": 9614034.0,
+ "step": 3940
+ },
+ {
+ "epoch": 9.494571773220748,
+ "eval_entropy": 0.17779575903596503,
+ "eval_loss": 1.148206114768982,
+ "eval_mean_token_accuracy": 0.8378150037835154,
+ "eval_num_tokens": 9614034.0,
+ "eval_runtime": 100.7077,
+ "eval_samples_per_second": 14.1,
+ "eval_steps_per_second": 1.767,
+ "step": 3940
+ },
+ {
+ "entropy": 0.08808465488255024,
+ "epoch": 9.54282267792521,
+ "grad_norm": 0.11133091151714325,
+ "learning_rate": 2.7837992240874838e-06,
+ "loss": 0.059873557090759276,
+ "mean_token_accuracy": 0.9792504802346229,
+ "num_tokens": 9661266.0,
+ "step": 3960
+ },
+ {
+ "epoch": 9.54282267792521,
+ "eval_entropy": 0.17780916221188695,
+ "eval_loss": 1.148876667022705,
+ "eval_mean_token_accuracy": 0.8378731015692936,
+ "eval_num_tokens": 9661266.0,
+ "eval_runtime": 100.8435,
+ "eval_samples_per_second": 14.081,
+ "eval_steps_per_second": 1.765,
+ "step": 3960
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 2.00357045156352e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4120/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4120/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4120/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4120/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4120/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4120/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4120/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4120/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4120/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4120/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4120/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4120/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4120/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4120/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..6cbdb7784cf8158c68358e13da48df136e0d1547
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4120/trainer_state.json
@@ -0,0 +1,4360 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 9.928829915560916,
+ "eval_steps": 20,
+ "global_step": 4120,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ },
+ {
+ "entropy": 0.6141144648194313,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6467089056968689,
+ "learning_rate": 7.078840074092574e-05,
+ "loss": 0.5548617362976074,
+ "mean_token_accuracy": 0.8342376738786698,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.5951976082967908,
+ "eval_loss": 0.5850171446800232,
+ "eval_mean_token_accuracy": 0.8303041870004675,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 98.6601,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 760
+ },
+ {
+ "entropy": 0.5961867064237595,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.41280242800712585,
+ "learning_rate": 7.061001336495465e-05,
+ "loss": 0.5511267185211182,
+ "mean_token_accuracy": 0.8392465204000473,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6167017998320333,
+ "eval_loss": 0.5826791524887085,
+ "eval_mean_token_accuracy": 0.8309678050239434,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 98.6564,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 780
+ },
+ {
+ "entropy": 0.6025628983974457,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.4078947901725769,
+ "learning_rate": 7.042187293157881e-05,
+ "loss": 0.5552026271820069,
+ "mean_token_accuracy": 0.8359778299927711,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.6047008484601974,
+ "eval_loss": 0.5811123847961426,
+ "eval_mean_token_accuracy": 0.8320730804057603,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 98.6033,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 800
+ },
+ {
+ "entropy": 0.6041995719075203,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.543890118598938,
+ "learning_rate": 7.022403268231535e-05,
+ "loss": 0.554384422302246,
+ "mean_token_accuracy": 0.8380729153752327,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.6005127917179901,
+ "eval_loss": 0.5800131559371948,
+ "eval_mean_token_accuracy": 0.8313186547059691,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 98.4119,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 820
+ },
+ {
+ "entropy": 0.567442889396961,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.5015465021133423,
+ "learning_rate": 7.001654860361456e-05,
+ "loss": 0.5208251476287842,
+ "mean_token_accuracy": 0.8442784960453327,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.5785037809878253,
+ "eval_loss": 0.5865591168403625,
+ "eval_mean_token_accuracy": 0.8307993857378371,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 98.5061,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 840
+ },
+ {
+ "entropy": 0.5552012033760547,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.5975943207740784,
+ "learning_rate": 6.979947941101646e-05,
+ "loss": 0.5064489841461182,
+ "mean_token_accuracy": 0.8468396887183189,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5844205239515626,
+ "eval_loss": 0.585284948348999,
+ "eval_mean_token_accuracy": 0.8309797065311604,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 98.4982,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 860
+ },
+ {
+ "entropy": 0.5401638634502888,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.4194582998752594,
+ "learning_rate": 6.957288653253486e-05,
+ "loss": 0.49037814140319824,
+ "mean_token_accuracy": 0.8528496414422989,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.574507936835289,
+ "eval_loss": 0.5836747288703918,
+ "eval_mean_token_accuracy": 0.8325241643391298,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 98.5592,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 880
+ },
+ {
+ "entropy": 0.5606910541653634,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5395355820655823,
+ "learning_rate": 6.933683409127401e-05,
+ "loss": 0.5089571475982666,
+ "mean_token_accuracy": 0.8459660410881042,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5779740023479033,
+ "eval_loss": 0.5816043615341187,
+ "eval_mean_token_accuracy": 0.8326318036304431,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 98.4386,
+ "eval_samples_per_second": 14.425,
+ "eval_steps_per_second": 1.808,
+ "step": 900
+ },
+ {
+ "entropy": 0.5380903735756875,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.46588417887687683,
+ "learning_rate": 6.909138888728257e-05,
+ "loss": 0.4902662754058838,
+ "mean_token_accuracy": 0.8493925213813782,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.570154261220707,
+ "eval_loss": 0.5827520489692688,
+ "eval_mean_token_accuracy": 0.8321914167216654,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 98.5674,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 920
+ },
+ {
+ "entropy": 0.558555257320404,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.44576048851013184,
+ "learning_rate": 6.883662037864996e-05,
+ "loss": 0.5000699520111084,
+ "mean_token_accuracy": 0.8475695163011551,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5715417749761196,
+ "eval_loss": 0.5831062197685242,
+ "eval_mean_token_accuracy": 0.8327409899636601,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 98.568,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 940
+ },
+ {
+ "entropy": 0.5582522705197335,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.38966405391693115,
+ "learning_rate": 6.857260066185045e-05,
+ "loss": 0.5132341861724854,
+ "mean_token_accuracy": 0.8466864034533501,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5823918579334624,
+ "eval_loss": 0.5797588229179382,
+ "eval_mean_token_accuracy": 0.8326037566313583,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 98.6222,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 960
+ },
+ {
+ "entropy": 0.5557353265583516,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.42300811409950256,
+ "learning_rate": 6.829940445134089e-05,
+ "loss": 0.5035375118255615,
+ "mean_token_accuracy": 0.8479988545179367,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5542296482605881,
+ "eval_loss": 0.5800604820251465,
+ "eval_mean_token_accuracy": 0.8336404163515969,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 98.615,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 980
+ },
+ {
+ "entropy": 0.5492161035537719,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.47482752799987793,
+ "learning_rate": 6.801710905841724e-05,
+ "loss": 0.5032052040100098,
+ "mean_token_accuracy": 0.8488596931099892,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5755258423558781,
+ "eval_loss": 0.5764774680137634,
+ "eval_mean_token_accuracy": 0.8336571763070781,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 98.4765,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5439704649150372,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3656543493270874,
+ "learning_rate": 6.772579436933644e-05,
+ "loss": 0.49903340339660646,
+ "mean_token_accuracy": 0.8499931409955025,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5712662409530597,
+ "eval_loss": 0.5777108073234558,
+ "eval_mean_token_accuracy": 0.8336941352051296,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 98.678,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5448883637785912,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.3990268111228943,
+ "learning_rate": 6.742554282270959e-05,
+ "loss": 0.4990490436553955,
+ "mean_token_accuracy": 0.8488436549901962,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5816385483139017,
+ "eval_loss": 0.5749375820159912,
+ "eval_mean_token_accuracy": 0.833801462074344,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 98.6209,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5575303137302399,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.45015832781791687,
+ "learning_rate": 6.711643938617279e-05,
+ "loss": 0.5061063289642334,
+ "mean_token_accuracy": 0.8462719902396202,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5666238901655326,
+ "eval_loss": 0.574820876121521,
+ "eval_mean_token_accuracy": 0.8339746617869045,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 98.5332,
+ "eval_samples_per_second": 14.411,
+ "eval_steps_per_second": 1.806,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5477629877626896,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.43583670258522034,
+ "learning_rate": 6.679857153234218e-05,
+ "loss": 0.4998013496398926,
+ "mean_token_accuracy": 0.8465577483177185,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5809191281206152,
+ "eval_loss": 0.5737497210502625,
+ "eval_mean_token_accuracy": 0.8338909467284599,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 98.5093,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5548589795827865,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.44145169854164124,
+ "learning_rate": 6.647202921406043e-05,
+ "loss": 0.5035099029541016,
+ "mean_token_accuracy": 0.8463503196835518,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5626686118292005,
+ "eval_loss": 0.5707846879959106,
+ "eval_mean_token_accuracy": 0.8349729531936432,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 98.473,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5568635657429695,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.38482189178466797,
+ "learning_rate": 6.613690483894096e-05,
+ "loss": 0.5090720176696777,
+ "mean_token_accuracy": 0.8481037884950637,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5752792709998871,
+ "eval_loss": 0.5706034898757935,
+ "eval_mean_token_accuracy": 0.8350465498613507,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 98.6524,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5644363485276699,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.4353690445423126,
+ "learning_rate": 6.57932932432178e-05,
+ "loss": 0.5078121185302734,
+ "mean_token_accuracy": 0.846393559873104,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5670730096570561,
+ "eval_loss": 0.5699080228805542,
+ "eval_mean_token_accuracy": 0.8354134288396728,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 98.5697,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 1140
+ },
+ {
+ "entropy": 0.547992604970932,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.4126238524913788,
+ "learning_rate": 6.544129166490802e-05,
+ "loss": 0.4943515777587891,
+ "mean_token_accuracy": 0.8496095523238182,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5552005295673114,
+ "eval_loss": 0.5671029090881348,
+ "eval_mean_token_accuracy": 0.8364275894807965,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 98.5412,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 1160
+ },
+ {
+ "entropy": 0.561221782118082,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.47753170132637024,
+ "learning_rate": 6.508099971629446e-05,
+ "loss": 0.5094280242919922,
+ "mean_token_accuracy": 0.8462468951940536,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5532578974962234,
+ "eval_loss": 0.5714129209518433,
+ "eval_mean_token_accuracy": 0.8357450737712089,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 98.5358,
+ "eval_samples_per_second": 14.411,
+ "eval_steps_per_second": 1.806,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5400963060557842,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.38486021757125854,
+ "learning_rate": 6.471251935573679e-05,
+ "loss": 0.49443836212158204,
+ "mean_token_accuracy": 0.8484652638435364,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.5697383115465722,
+ "eval_loss": 0.5681865215301514,
+ "eval_mean_token_accuracy": 0.8350923453154189,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 98.4136,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5381059750914574,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.49116969108581543,
+ "learning_rate": 6.43359548588183e-05,
+ "loss": 0.4878206729888916,
+ "mean_token_accuracy": 0.8522022038698196,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5614734472853414,
+ "eval_loss": 0.5665943622589111,
+ "eval_mean_token_accuracy": 0.8360684296388304,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 98.5682,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5569418989121914,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.39628803730010986,
+ "learning_rate": 6.39514127888373e-05,
+ "loss": 0.5087827682495117,
+ "mean_token_accuracy": 0.8476181954145432,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5740904657358534,
+ "eval_loss": 0.5637108683586121,
+ "eval_mean_token_accuracy": 0.8366153638684348,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 98.4904,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 1240
+ },
+ {
+ "entropy": 0.5067426371268737,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.4593591094017029,
+ "learning_rate": 6.355900196665094e-05,
+ "loss": 0.4515994071960449,
+ "mean_token_accuracy": 0.8621168640943674,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5176989454566763,
+ "eval_loss": 0.5825926661491394,
+ "eval_mean_token_accuracy": 0.8353930775369152,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 98.5654,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 1260
+ },
+ {
+ "entropy": 0.4712781891226768,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.46717822551727295,
+ "learning_rate": 6.315883343988008e-05,
+ "loss": 0.4228676795959473,
+ "mean_token_accuracy": 0.8672123551368713,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.5229383730486538,
+ "eval_loss": 0.5875843167304993,
+ "eval_mean_token_accuracy": 0.8347370524084969,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 98.5459,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 1280
+ },
+ {
+ "entropy": 0.48235836550593375,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.5073364973068237,
+ "learning_rate": 6.275102045148425e-05,
+ "loss": 0.4365375518798828,
+ "mean_token_accuracy": 0.8644895792007447,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.5151654947339819,
+ "eval_loss": 0.5826129913330078,
+ "eval_mean_token_accuracy": 0.8359152366606037,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 98.3713,
+ "eval_samples_per_second": 14.435,
+ "eval_steps_per_second": 1.809,
+ "step": 1300
+ },
+ {
+ "entropy": 0.4756111137568951,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.4896017909049988,
+ "learning_rate": 6.233567840771526e-05,
+ "loss": 0.43177642822265627,
+ "mean_token_accuracy": 0.8645799145102501,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5227621813838401,
+ "eval_loss": 0.5796895027160645,
+ "eval_mean_token_accuracy": 0.8355028448479899,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 98.5538,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 1320
+ },
+ {
+ "entropy": 0.47541345208883284,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5064273476600647,
+ "learning_rate": 6.191292484545842e-05,
+ "loss": 0.4289750099182129,
+ "mean_token_accuracy": 0.8659383848309516,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.5286844237132019,
+ "eval_loss": 0.579947829246521,
+ "eval_mean_token_accuracy": 0.8353635648663125,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 98.589,
+ "eval_samples_per_second": 14.403,
+ "eval_steps_per_second": 1.805,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4916382558643818,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.5225116014480591,
+ "learning_rate": 6.148287939897118e-05,
+ "loss": 0.44352192878723146,
+ "mean_token_accuracy": 0.8630174115300179,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.5264859509267165,
+ "eval_loss": 0.5774893164634705,
+ "eval_mean_token_accuracy": 0.8363003827882617,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 98.5858,
+ "eval_samples_per_second": 14.404,
+ "eval_steps_per_second": 1.806,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4990046709775925,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.548435628414154,
+ "learning_rate": 6.104566376602795e-05,
+ "loss": 0.4463675498962402,
+ "mean_token_accuracy": 0.8630342438817025,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.522659856616781,
+ "eval_loss": 0.5790286660194397,
+ "eval_mean_token_accuracy": 0.8355394277010071,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 98.4707,
+ "eval_samples_per_second": 14.421,
+ "eval_steps_per_second": 1.808,
+ "step": 1380
+ },
+ {
+ "entropy": 0.48943497613072395,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6222821474075317,
+ "learning_rate": 6.060140167348109e-05,
+ "loss": 0.4356812000274658,
+ "mean_token_accuracy": 0.8644352272152901,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.5065617703654793,
+ "eval_loss": 0.5842992067337036,
+ "eval_mean_token_accuracy": 0.8358401662847968,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 98.4775,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 1400
+ },
+ {
+ "entropy": 0.47648413479328156,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.5207330584526062,
+ "learning_rate": 6.015021884224766e-05,
+ "loss": 0.42724952697753904,
+ "mean_token_accuracy": 0.8663465172052384,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.5232780023237292,
+ "eval_loss": 0.5769163966178894,
+ "eval_mean_token_accuracy": 0.8356800042511372,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 98.5005,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4767202906310558,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4695364534854889,
+ "learning_rate": 5.969224295173193e-05,
+ "loss": 0.42672252655029297,
+ "mean_token_accuracy": 0.8657647699117661,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.5146279760291067,
+ "eval_loss": 0.577904224395752,
+ "eval_mean_token_accuracy": 0.8365512393163831,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 98.5825,
+ "eval_samples_per_second": 14.404,
+ "eval_steps_per_second": 1.806,
+ "step": 1440
+ },
+ {
+ "entropy": 0.4855267658829689,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6333373785018921,
+ "learning_rate": 5.92276036036936e-05,
+ "loss": 0.4310962677001953,
+ "mean_token_accuracy": 0.864738130569458,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5260473481389913,
+ "eval_loss": 0.5770025253295898,
+ "eval_mean_token_accuracy": 0.8357346339172191,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 98.462,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 1460
+ },
+ {
+ "entropy": 0.4917050123214722,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.5833342671394348,
+ "learning_rate": 5.875643228557199e-05,
+ "loss": 0.44533047676086424,
+ "mean_token_accuracy": 0.8642907112836837,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.5278292167722509,
+ "eval_loss": 0.5773910880088806,
+ "eval_mean_token_accuracy": 0.83580608548743,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 98.4817,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 1480
+ },
+ {
+ "entropy": 0.5025837384164333,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.5484253168106079,
+ "learning_rate": 5.827886233327671e-05,
+ "loss": 0.44471564292907717,
+ "mean_token_accuracy": 0.8624475196003913,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.5118448720889145,
+ "eval_loss": 0.5786464810371399,
+ "eval_mean_token_accuracy": 0.8370849305994055,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 98.5208,
+ "eval_samples_per_second": 14.413,
+ "eval_steps_per_second": 1.807,
+ "step": 1500
+ },
+ {
+ "entropy": 0.494297581911087,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5475512146949768,
+ "learning_rate": 5.7795028893455165e-05,
+ "loss": 0.44118366241455076,
+ "mean_token_accuracy": 0.8630154177546501,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.5250470112548785,
+ "eval_loss": 0.5746602416038513,
+ "eval_mean_token_accuracy": 0.8370710711800651,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 98.54,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 1520
+ },
+ {
+ "entropy": 0.4840923048555851,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.39450952410697937,
+ "learning_rate": 5.730506888524761e-05,
+ "loss": 0.443526554107666,
+ "mean_token_accuracy": 0.8627525731921196,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5305671740448876,
+ "eval_loss": 0.5697008371353149,
+ "eval_mean_token_accuracy": 0.837059904015466,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 98.6288,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 1540
+ },
+ {
+ "entropy": 0.49613324701786043,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.47117286920547485,
+ "learning_rate": 5.680912096154071e-05,
+ "loss": 0.44313778877258303,
+ "mean_token_accuracy": 0.8614323407411575,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.522762700580479,
+ "eval_loss": 0.571703314781189,
+ "eval_mean_token_accuracy": 0.8381530970669864,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 98.6493,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4979200787842274,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.5275628566741943,
+ "learning_rate": 5.630732546973038e-05,
+ "loss": 0.4462272167205811,
+ "mean_token_accuracy": 0.8615201056003571,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.5232290047942922,
+ "eval_loss": 0.568954586982727,
+ "eval_mean_token_accuracy": 0.8380421128835571,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 98.484,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4957688055932522,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5806206464767456,
+ "learning_rate": 5.5799824412005126e-05,
+ "loss": 0.44518418312072755,
+ "mean_token_accuracy": 0.860895662009716,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.5238539929805177,
+ "eval_loss": 0.573081910610199,
+ "eval_mean_token_accuracy": 0.8369396288743179,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 98.5731,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 1600
+ },
+ {
+ "entropy": 0.5005358196794987,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.5114495754241943,
+ "learning_rate": 5.528676140516111e-05,
+ "loss": 0.45078649520874026,
+ "mean_token_accuracy": 0.860628567636013,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.5245393164707034,
+ "eval_loss": 0.5683825612068176,
+ "eval_mean_token_accuracy": 0.8380641599049728,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 98.5275,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 1620
+ },
+ {
+ "entropy": 0.4839292488992214,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6169933080673218,
+ "learning_rate": 5.4768281639960344e-05,
+ "loss": 0.43407382965087893,
+ "mean_token_accuracy": 0.867341585457325,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.5277225765284528,
+ "eval_loss": 0.567344069480896,
+ "eval_mean_token_accuracy": 0.8386145431673928,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 98.5702,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 1640
+ },
+ {
+ "entropy": 0.4952052327302786,
+ "epoch": 4.0,
+ "grad_norm": 2.080777168273926,
+ "learning_rate": 5.424453184004336e-05,
+ "loss": 0.4525449752807617,
+ "mean_token_accuracy": 0.8622837005517422,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.5197472232446242,
+ "eval_loss": 0.57077956199646,
+ "eval_mean_token_accuracy": 0.8381094560864266,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 98.5498,
+ "eval_samples_per_second": 14.409,
+ "eval_steps_per_second": 1.806,
+ "step": 1660
+ },
+ {
+ "entropy": 0.42416453957557676,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.41175922751426697,
+ "learning_rate": 5.3715660220408146e-05,
+ "loss": 0.3610882043838501,
+ "mean_token_accuracy": 0.8832842499017716,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.4814340365066957,
+ "eval_loss": 0.5963674783706665,
+ "eval_mean_token_accuracy": 0.8364584405770462,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 98.4464,
+ "eval_samples_per_second": 14.424,
+ "eval_steps_per_second": 1.808,
+ "step": 1680
+ },
+ {
+ "entropy": 0.4193890765309334,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.4773254990577698,
+ "learning_rate": 5.3181816445467126e-05,
+ "loss": 0.36023797988891604,
+ "mean_token_accuracy": 0.884614959359169,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.47795870448096417,
+ "eval_loss": 0.6029345393180847,
+ "eval_mean_token_accuracy": 0.8354013039824668,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 98.5784,
+ "eval_samples_per_second": 14.405,
+ "eval_steps_per_second": 1.806,
+ "step": 1700
+ },
+ {
+ "entropy": 0.41848619729280473,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.7293462157249451,
+ "learning_rate": 5.264315158669384e-05,
+ "loss": 0.36621780395507814,
+ "mean_token_accuracy": 0.882377204298973,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4683974619661824,
+ "eval_loss": 0.6010014414787292,
+ "eval_mean_token_accuracy": 0.8358984790491254,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 98.5406,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 1720
+ },
+ {
+ "entropy": 0.41377718672156333,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5061255693435669,
+ "learning_rate": 5.209981807987152e-05,
+ "loss": 0.36024787425994875,
+ "mean_token_accuracy": 0.8844522640109063,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.4666026450609893,
+ "eval_loss": 0.6050319075584412,
+ "eval_mean_token_accuracy": 0.8357557922266843,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 98.6056,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 1740
+ },
+ {
+ "entropy": 0.4185848042368889,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.49589282274246216,
+ "learning_rate": 5.155196968195561e-05,
+ "loss": 0.36044313907623293,
+ "mean_token_accuracy": 0.8863146558403969,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.4660757115382827,
+ "eval_loss": 0.6089843511581421,
+ "eval_mean_token_accuracy": 0.8352006492989786,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 98.5425,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 1760
+ },
+ {
+ "entropy": 0.4249664135277271,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.6652432680130005,
+ "learning_rate": 5.099976142756227e-05,
+ "loss": 0.37266385555267334,
+ "mean_token_accuracy": 0.880594964325428,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.48157174483443915,
+ "eval_loss": 0.5987246036529541,
+ "eval_mean_token_accuracy": 0.8361060595914219,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 98.4652,
+ "eval_samples_per_second": 14.421,
+ "eval_steps_per_second": 1.808,
+ "step": 1780
+ },
+ {
+ "entropy": 0.4195803053677082,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.6217629313468933,
+ "learning_rate": 5.044334958509551e-05,
+ "loss": 0.3646178483963013,
+ "mean_token_accuracy": 0.8808894723653793,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.47762840182593697,
+ "eval_loss": 0.5994009375572205,
+ "eval_mean_token_accuracy": 0.8366789422678144,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 98.6677,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 1800
+ },
+ {
+ "entropy": 0.42473565861582757,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.6135475635528564,
+ "learning_rate": 4.9882891612525023e-05,
+ "loss": 0.37249753475189207,
+ "mean_token_accuracy": 0.8800962314009666,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.47359335924802204,
+ "eval_loss": 0.6024664640426636,
+ "eval_mean_token_accuracy": 0.8355274846714534,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 98.5103,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 1820
+ },
+ {
+ "entropy": 0.4212189674377441,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.5874977707862854,
+ "learning_rate": 4.931854611282748e-05,
+ "loss": 0.36766781806945803,
+ "mean_token_accuracy": 0.8837772741913795,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.46670635798004234,
+ "eval_loss": 0.6085236668586731,
+ "eval_mean_token_accuracy": 0.8355892652206207,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 98.5871,
+ "eval_samples_per_second": 14.404,
+ "eval_steps_per_second": 1.806,
+ "step": 1840
+ },
+ {
+ "entropy": 0.41604730263352396,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5763360261917114,
+ "learning_rate": 4.875047278910371e-05,
+ "loss": 0.36643099784851074,
+ "mean_token_accuracy": 0.8825299561023712,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.47700903121005284,
+ "eval_loss": 0.606238842010498,
+ "eval_mean_token_accuracy": 0.8348092550642034,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 98.5305,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 1860
+ },
+ {
+ "entropy": 0.4189125746488571,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5916087627410889,
+ "learning_rate": 4.817883239938471e-05,
+ "loss": 0.3703711748123169,
+ "mean_token_accuracy": 0.8825044095516205,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.47982015703501324,
+ "eval_loss": 0.6003754138946533,
+ "eval_mean_token_accuracy": 0.8351969742373134,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 98.4111,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 1880
+ },
+ {
+ "entropy": 0.4429341182112694,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.6441562175750732,
+ "learning_rate": 4.7603786711138866e-05,
+ "loss": 0.3824815511703491,
+ "mean_token_accuracy": 0.8784194126725197,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.4740966230296017,
+ "eval_loss": 0.5988267660140991,
+ "eval_mean_token_accuracy": 0.8369577316085944,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 98.5291,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 1900
+ },
+ {
+ "entropy": 0.41137003526091576,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.6663628816604614,
+ "learning_rate": 4.702549845549385e-05,
+ "loss": 0.36178064346313477,
+ "mean_token_accuracy": 0.8830229461193084,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.4730259802904022,
+ "eval_loss": 0.6001402139663696,
+ "eval_mean_token_accuracy": 0.8366120139534554,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 98.5556,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 1920
+ },
+ {
+ "entropy": 0.4264552935957909,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6708512902259827,
+ "learning_rate": 4.64441312811855e-05,
+ "loss": 0.376303505897522,
+ "mean_token_accuracy": 0.8826978862285614,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4687648815720269,
+ "eval_loss": 0.5982367992401123,
+ "eval_mean_token_accuracy": 0.8368918728292658,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 98.4928,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 1940
+ },
+ {
+ "entropy": 0.4378764644265175,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.7098956108093262,
+ "learning_rate": 4.585984970824725e-05,
+ "loss": 0.38264145851135256,
+ "mean_token_accuracy": 0.8768767058849335,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.46603381248672354,
+ "eval_loss": 0.6010599732398987,
+ "eval_mean_token_accuracy": 0.8361037319965576,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 98.5157,
+ "eval_samples_per_second": 14.414,
+ "eval_steps_per_second": 1.807,
+ "step": 1960
+ },
+ {
+ "entropy": 0.44009693413972856,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.7529758214950562,
+ "learning_rate": 4.527281908145284e-05,
+ "loss": 0.379723048210144,
+ "mean_token_accuracy": 0.8799900904297828,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.47042709873633437,
+ "eval_loss": 0.5984888672828674,
+ "eval_mean_token_accuracy": 0.8375813492228475,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 98.3352,
+ "eval_samples_per_second": 14.44,
+ "eval_steps_per_second": 1.81,
+ "step": 1980
+ },
+ {
+ "entropy": 0.43963938802480695,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.530246376991272,
+ "learning_rate": 4.468320552352577e-05,
+ "loss": 0.3855365037918091,
+ "mean_token_accuracy": 0.8766040325164794,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4773505919435051,
+ "eval_loss": 0.5956445932388306,
+ "eval_mean_token_accuracy": 0.8365653522228926,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 98.6015,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 2000
+ },
+ {
+ "entropy": 0.4241503834724426,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.5264645218849182,
+ "learning_rate": 4.409117588812845e-05,
+ "loss": 0.3725905418395996,
+ "mean_token_accuracy": 0.8816388875246048,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4738255136468437,
+ "eval_loss": 0.5987666845321655,
+ "eval_mean_token_accuracy": 0.8362959995028678,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 98.6141,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 2020
+ },
+ {
+ "entropy": 0.4451489016413689,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.5322316884994507,
+ "learning_rate": 4.3496897712644684e-05,
+ "loss": 0.3903519630432129,
+ "mean_token_accuracy": 0.8752380400896073,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.4662560015247109,
+ "eval_loss": 0.5963740944862366,
+ "eval_mean_token_accuracy": 0.8378168083308788,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 98.6613,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 2040
+ },
+ {
+ "entropy": 0.43658976554870604,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.6989460587501526,
+ "learning_rate": 4.29005391707685e-05,
+ "loss": 0.3851702451705933,
+ "mean_token_accuracy": 0.8790383815765381,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4791082282749455,
+ "eval_loss": 0.5912667512893677,
+ "eval_mean_token_accuracy": 0.83739292286755,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 98.5606,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 2060
+ },
+ {
+ "entropy": 0.40598615316244274,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.5303810238838196,
+ "learning_rate": 4.2302269024913e-05,
+ "loss": 0.3450628280639648,
+ "mean_token_accuracy": 0.8893755567379487,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.42836171187711564,
+ "eval_loss": 0.6335396766662598,
+ "eval_mean_token_accuracy": 0.8349329397249757,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 98.3503,
+ "eval_samples_per_second": 14.438,
+ "eval_steps_per_second": 1.81,
+ "step": 2080
+ },
+ {
+ "entropy": 0.3517624609172344,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.6402326226234436,
+ "learning_rate": 4.170225657845261e-05,
+ "loss": 0.2898552894592285,
+ "mean_token_accuracy": 0.904600378870964,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.4094204810563098,
+ "eval_loss": 0.652080237865448,
+ "eval_mean_token_accuracy": 0.8331165705504042,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 98.3573,
+ "eval_samples_per_second": 14.437,
+ "eval_steps_per_second": 1.81,
+ "step": 2100
+ },
+ {
+ "entropy": 0.3560003057122231,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.9274800419807434,
+ "learning_rate": 4.110067162781221e-05,
+ "loss": 0.2926265954971313,
+ "mean_token_accuracy": 0.9040415585041046,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.41318756994906436,
+ "eval_loss": 0.6567395925521851,
+ "eval_mean_token_accuracy": 0.83362964197491,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 98.5042,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 2120
+ },
+ {
+ "entropy": 0.35080533251166346,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.6713427901268005,
+ "learning_rate": 4.049768441441677e-05,
+ "loss": 0.29822506904602053,
+ "mean_token_accuracy": 0.9027505293488503,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.416848015416874,
+ "eval_loss": 0.6563553214073181,
+ "eval_mean_token_accuracy": 0.8333374677079447,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 98.4261,
+ "eval_samples_per_second": 14.427,
+ "eval_steps_per_second": 1.808,
+ "step": 2140
+ },
+ {
+ "entropy": 0.3736354000866413,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.5667551755905151,
+ "learning_rate": 3.989346557651503e-05,
+ "loss": 0.3088909387588501,
+ "mean_token_accuracy": 0.9012129113078118,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.41974830309326727,
+ "eval_loss": 0.6504881381988525,
+ "eval_mean_token_accuracy": 0.8349264759695931,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 98.5007,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 2160
+ },
+ {
+ "entropy": 0.35063633099198344,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.7737690210342407,
+ "learning_rate": 3.9288186100890903e-05,
+ "loss": 0.2888066053390503,
+ "mean_token_accuracy": 0.9040960729122162,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.4058936465992017,
+ "eval_loss": 0.6578952670097351,
+ "eval_mean_token_accuracy": 0.8352396407823884,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 98.4807,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 2180
+ },
+ {
+ "entropy": 0.3545838728547096,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.8293929100036621,
+ "learning_rate": 3.8682017274476215e-05,
+ "loss": 0.29347708225250246,
+ "mean_token_accuracy": 0.902638104557991,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.41738414345832353,
+ "eval_loss": 0.6543260812759399,
+ "eval_mean_token_accuracy": 0.8336448260907376,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 98.499,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 2200
+ },
+ {
+ "entropy": 0.3416903533041477,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.7580088973045349,
+ "learning_rate": 3.8075130635878556e-05,
+ "loss": 0.28760058879852296,
+ "mean_token_accuracy": 0.906610831618309,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.4121073748958245,
+ "eval_loss": 0.6563934087753296,
+ "eval_mean_token_accuracy": 0.833708616120092,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 98.3808,
+ "eval_samples_per_second": 14.434,
+ "eval_steps_per_second": 1.809,
+ "step": 2220
+ },
+ {
+ "entropy": 0.352034630253911,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.7154259085655212,
+ "learning_rate": 3.746769792683787e-05,
+ "loss": 0.2967940092086792,
+ "mean_token_accuracy": 0.9023159235715866,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.4149914874454563,
+ "eval_loss": 0.6444063186645508,
+ "eval_mean_token_accuracy": 0.8360218017288809,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 98.4832,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 2240
+ },
+ {
+ "entropy": 0.3539439059793949,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.8078773617744446,
+ "learning_rate": 3.685989104362558e-05,
+ "loss": 0.2928986310958862,
+ "mean_token_accuracy": 0.905618867278099,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.4099769044793054,
+ "eval_loss": 0.6588519215583801,
+ "eval_mean_token_accuracy": 0.8353862765799748,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 98.3781,
+ "eval_samples_per_second": 14.434,
+ "eval_steps_per_second": 1.809,
+ "step": 2260
+ },
+ {
+ "entropy": 0.35821759328246117,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.8760809302330017,
+ "learning_rate": 3.6251881988399994e-05,
+ "loss": 0.29706525802612305,
+ "mean_token_accuracy": 0.9018172711133957,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.40791591667057425,
+ "eval_loss": 0.6583756804466248,
+ "eval_mean_token_accuracy": 0.8349186125096311,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 98.4808,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 2280
+ },
+ {
+ "entropy": 0.34672123566269875,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.8768810033798218,
+ "learning_rate": 3.564384282053165e-05,
+ "loss": 0.2885677576065063,
+ "mean_token_accuracy": 0.9063566774129868,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.412923522377282,
+ "eval_loss": 0.6535082459449768,
+ "eval_mean_token_accuracy": 0.8350685978873392,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 98.3842,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 2300
+ },
+ {
+ "entropy": 0.35606732666492463,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.7237412929534912,
+ "learning_rate": 3.503594560791267e-05,
+ "loss": 0.30418403148651124,
+ "mean_token_accuracy": 0.9004160806536674,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.42160589155856143,
+ "eval_loss": 0.6412338018417358,
+ "eval_mean_token_accuracy": 0.834980196162556,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 98.3674,
+ "eval_samples_per_second": 14.436,
+ "eval_steps_per_second": 1.81,
+ "step": 2320
+ },
+ {
+ "entropy": 0.35330103561282156,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.5864736437797546,
+ "learning_rate": 3.442836237826349e-05,
+ "loss": 0.29608047008514404,
+ "mean_token_accuracy": 0.9029312193393707,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.4273778943533308,
+ "eval_loss": 0.6407549977302551,
+ "eval_mean_token_accuracy": 0.8351332397943132,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 98.5533,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 2340
+ },
+ {
+ "entropy": 0.353700140863657,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.744305431842804,
+ "learning_rate": 3.382126507045109e-05,
+ "loss": 0.3004793643951416,
+ "mean_token_accuracy": 0.9015571728348732,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.4262272163053577,
+ "eval_loss": 0.6443405151367188,
+ "eval_mean_token_accuracy": 0.8354629797212193,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 98.3719,
+ "eval_samples_per_second": 14.435,
+ "eval_steps_per_second": 1.809,
+ "step": 2360
+ },
+ {
+ "entropy": 0.36056690439581873,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.6552110910415649,
+ "learning_rate": 3.321482548583228e-05,
+ "loss": 0.2950709104537964,
+ "mean_token_accuracy": 0.9033273816108703,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.41902332275771026,
+ "eval_loss": 0.6472867727279663,
+ "eval_mean_token_accuracy": 0.8354150207524889,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 98.3956,
+ "eval_samples_per_second": 14.432,
+ "eval_steps_per_second": 1.809,
+ "step": 2380
+ },
+ {
+ "entropy": 0.35097662582993505,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.670831024646759,
+ "learning_rate": 3.260921523963606e-05,
+ "loss": 0.29224560260772703,
+ "mean_token_accuracy": 0.9042141750454903,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.4284254325909561,
+ "eval_loss": 0.6407246589660645,
+ "eval_mean_token_accuracy": 0.8345498831084605,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 98.3651,
+ "eval_samples_per_second": 14.436,
+ "eval_steps_per_second": 1.81,
+ "step": 2400
+ },
+ {
+ "entropy": 0.35357567593455314,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.8287181258201599,
+ "learning_rate": 3.200460571239839e-05,
+ "loss": 0.2948127746582031,
+ "mean_token_accuracy": 0.9042218685150146,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.4094803942723221,
+ "eval_loss": 0.6497060060501099,
+ "eval_mean_token_accuracy": 0.8358801147910986,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 98.3681,
+ "eval_samples_per_second": 14.436,
+ "eval_steps_per_second": 1.81,
+ "step": 2420
+ },
+ {
+ "entropy": 0.36021770909428596,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.6800532937049866,
+ "learning_rate": 3.140116800146367e-05,
+ "loss": 0.30136709213256835,
+ "mean_token_accuracy": 0.9020729199051857,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.4202987858083811,
+ "eval_loss": 0.6428810954093933,
+ "eval_mean_token_accuracy": 0.8355459637856215,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 98.5147,
+ "eval_samples_per_second": 14.414,
+ "eval_steps_per_second": 1.807,
+ "step": 2440
+ },
+ {
+ "entropy": 0.34706664383411406,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.6911236643791199,
+ "learning_rate": 3.07990728725661e-05,
+ "loss": 0.28953113555908205,
+ "mean_token_accuracy": 0.9055185422301293,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.406153663825453,
+ "eval_loss": 0.6541930437088013,
+ "eval_mean_token_accuracy": 0.834371502814668,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 98.3494,
+ "eval_samples_per_second": 14.438,
+ "eval_steps_per_second": 1.81,
+ "step": 2460
+ },
+ {
+ "entropy": 0.34852043315768244,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.7544664144515991,
+ "learning_rate": 3.0198490711505038e-05,
+ "loss": 0.28950369358062744,
+ "mean_token_accuracy": 0.9057677388191223,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.4174931427735961,
+ "eval_loss": 0.6448875665664673,
+ "eval_mean_token_accuracy": 0.8358313615402478,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 98.4786,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 2480
+ },
+ {
+ "entropy": 0.31319292042499935,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.7616425156593323,
+ "learning_rate": 2.9599591475927893e-05,
+ "loss": 0.24788825511932372,
+ "mean_token_accuracy": 0.9198053983541635,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.3601247634780541,
+ "eval_loss": 0.7335909008979797,
+ "eval_mean_token_accuracy": 0.8309332149082356,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 98.3569,
+ "eval_samples_per_second": 14.437,
+ "eval_steps_per_second": 1.81,
+ "step": 2500
+ },
+ {
+ "entropy": 0.27905523739755156,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.7366685271263123,
+ "learning_rate": 2.9002544647234015e-05,
+ "loss": 0.21279211044311525,
+ "mean_token_accuracy": 0.929847776889801,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.36974305969275784,
+ "eval_loss": 0.7080109715461731,
+ "eval_mean_token_accuracy": 0.8330447111236915,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 98.4908,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 2520
+ },
+ {
+ "entropy": 0.2782863024622202,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.9167613387107849,
+ "learning_rate": 2.840751918261356e-05,
+ "loss": 0.20897085666656495,
+ "mean_token_accuracy": 0.9294204145669938,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.366466105319141,
+ "eval_loss": 0.7170661091804504,
+ "eval_mean_token_accuracy": 0.8328661680891273,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 98.4385,
+ "eval_samples_per_second": 14.425,
+ "eval_steps_per_second": 1.808,
+ "step": 2540
+ },
+ {
+ "entropy": 0.28355731777846815,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.8010057806968689,
+ "learning_rate": 2.7814683467234538e-05,
+ "loss": 0.211204195022583,
+ "mean_token_accuracy": 0.9297966167330742,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.36755596988656547,
+ "eval_loss": 0.7274795174598694,
+ "eval_mean_token_accuracy": 0.8303602230682802,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 98.487,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 2560
+ },
+ {
+ "entropy": 0.27822012901306153,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.7516391277313232,
+ "learning_rate": 2.7224205266591812e-05,
+ "loss": 0.2117089033126831,
+ "mean_token_accuracy": 0.9294167309999466,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.37035224410925016,
+ "eval_loss": 0.7123831510543823,
+ "eval_mean_token_accuracy": 0.8318134563022785,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 98.5029,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 2580
+ },
+ {
+ "entropy": 0.2784307342022657,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.8481637239456177,
+ "learning_rate": 2.6636251679031508e-05,
+ "loss": 0.21476459503173828,
+ "mean_token_accuracy": 0.9293668136000633,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.3675622491354353,
+ "eval_loss": 0.7221675515174866,
+ "eval_mean_token_accuracy": 0.8314910557832611,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 98.3892,
+ "eval_samples_per_second": 14.432,
+ "eval_steps_per_second": 1.809,
+ "step": 2600
+ },
+ {
+ "entropy": 0.27380565330386164,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.7563941478729248,
+ "learning_rate": 2.605098908846406e-05,
+ "loss": 0.2117525815963745,
+ "mean_token_accuracy": 0.9299218639731407,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.3676499978879864,
+ "eval_loss": 0.7249981164932251,
+ "eval_mean_token_accuracy": 0.8315788115678209,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 98.491,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 2620
+ },
+ {
+ "entropy": 0.26819119676947595,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.9161278605461121,
+ "learning_rate": 2.546858311727963e-05,
+ "loss": 0.2020425319671631,
+ "mean_token_accuracy": 0.9322459831833839,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.3589264298422953,
+ "eval_loss": 0.737293004989624,
+ "eval_mean_token_accuracy": 0.8307386482699534,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 98.4767,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 2640
+ },
+ {
+ "entropy": 0.275980044901371,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.7255368828773499,
+ "learning_rate": 2.488919857947885e-05,
+ "loss": 0.21207733154296876,
+ "mean_token_accuracy": 0.9280627354979515,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.3729423410102223,
+ "eval_loss": 0.7137707471847534,
+ "eval_mean_token_accuracy": 0.8315478007445175,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 98.384,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 2660
+ },
+ {
+ "entropy": 0.2794003043323755,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.817340612411499,
+ "learning_rate": 2.4312999434032465e-05,
+ "loss": 0.2155231714248657,
+ "mean_token_accuracy": 0.930096672475338,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.36857131709543506,
+ "eval_loss": 0.7273393869400024,
+ "eval_mean_token_accuracy": 0.8302122903673836,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 98.5119,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 2680
+ },
+ {
+ "entropy": 0.28550401143729687,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.7945013642311096,
+ "learning_rate": 2.374014873848286e-05,
+ "loss": 0.21428213119506836,
+ "mean_token_accuracy": 0.9295192703604698,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.3720237272843886,
+ "eval_loss": 0.7190349698066711,
+ "eval_mean_token_accuracy": 0.8307539365264807,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 98.3972,
+ "eval_samples_per_second": 14.431,
+ "eval_steps_per_second": 1.809,
+ "step": 2700
+ },
+ {
+ "entropy": 0.2789166793227196,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.7820216417312622,
+ "learning_rate": 2.3170808602800694e-05,
+ "loss": 0.21065421104431153,
+ "mean_token_accuracy": 0.9293019056320191,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.36198576715555086,
+ "eval_loss": 0.7250529527664185,
+ "eval_mean_token_accuracy": 0.8310985079642093,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 98.516,
+ "eval_samples_per_second": 14.414,
+ "eval_steps_per_second": 1.807,
+ "step": 2720
+ },
+ {
+ "entropy": 0.2805891681462526,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.8589067459106445,
+ "learning_rate": 2.2605140143509734e-05,
+ "loss": 0.21532771587371827,
+ "mean_token_accuracy": 0.9301707550883294,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.36839111149311066,
+ "eval_loss": 0.720038890838623,
+ "eval_mean_token_accuracy": 0.8314180384191234,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 98.3864,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 2740
+ },
+ {
+ "entropy": 0.2923452850431204,
+ "epoch": 6.651387213510254,
+ "grad_norm": 1.1464570760726929,
+ "learning_rate": 2.2043303438092746e-05,
+ "loss": 0.22156639099121095,
+ "mean_token_accuracy": 0.9259255439043045,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.36308020152402726,
+ "eval_loss": 0.7283808588981628,
+ "eval_mean_token_accuracy": 0.8307237183110098,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 98.3733,
+ "eval_samples_per_second": 14.435,
+ "eval_steps_per_second": 1.809,
+ "step": 2760
+ },
+ {
+ "entropy": 0.2749974336475134,
+ "epoch": 6.699638118214716,
+ "grad_norm": 1.0338026285171509,
+ "learning_rate": 2.1485457479691453e-05,
+ "loss": 0.2111821413040161,
+ "mean_token_accuracy": 0.9308380514383316,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.35679193979568696,
+ "eval_loss": 0.7316031455993652,
+ "eval_mean_token_accuracy": 0.8316076066386834,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 98.4012,
+ "eval_samples_per_second": 14.431,
+ "eval_steps_per_second": 1.809,
+ "step": 2780
+ },
+ {
+ "entropy": 0.2827217370271683,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.6860962510108948,
+ "learning_rate": 2.0931760132113435e-05,
+ "loss": 0.219130802154541,
+ "mean_token_accuracy": 0.9272046208381652,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.367448085479522,
+ "eval_loss": 0.7210394144058228,
+ "eval_mean_token_accuracy": 0.8324049028118005,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 98.3867,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 2800
+ },
+ {
+ "entropy": 0.28139964006841184,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.7198498249053955,
+ "learning_rate": 2.038236808515841e-05,
+ "loss": 0.2179172992706299,
+ "mean_token_accuracy": 0.9282332047820091,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.3658287344353922,
+ "eval_loss": 0.7204482555389404,
+ "eval_mean_token_accuracy": 0.8324238242727987,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 98.3796,
+ "eval_samples_per_second": 14.434,
+ "eval_steps_per_second": 1.809,
+ "step": 2820
+ },
+ {
+ "entropy": 0.28035391345620153,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.8470256924629211,
+ "learning_rate": 1.9837436810277026e-05,
+ "loss": 0.21534478664398193,
+ "mean_token_accuracy": 0.9284071952104569,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.3615414013018769,
+ "eval_loss": 0.7223248481750488,
+ "eval_mean_token_accuracy": 0.8328172581249409,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 98.4426,
+ "eval_samples_per_second": 14.425,
+ "eval_steps_per_second": 1.808,
+ "step": 2840
+ },
+ {
+ "entropy": 0.27447608187794686,
+ "epoch": 6.892641737032569,
+ "grad_norm": 1.1274205446243286,
+ "learning_rate": 1.9297120516574132e-05,
+ "loss": 0.2151768684387207,
+ "mean_token_accuracy": 0.9284873619675637,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.36038200892089456,
+ "eval_loss": 0.7199941873550415,
+ "eval_mean_token_accuracy": 0.8334562922461649,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 98.4807,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 2860
+ },
+ {
+ "entropy": 0.27447244003415106,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.857891857624054,
+ "learning_rate": 1.876157210716952e-05,
+ "loss": 0.21279585361480713,
+ "mean_token_accuracy": 0.9296008720993996,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.3578391455197602,
+ "eval_loss": 0.7355349063873291,
+ "eval_mean_token_accuracy": 0.8319483108065101,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 98.3412,
+ "eval_samples_per_second": 14.44,
+ "eval_steps_per_second": 1.81,
+ "step": 2880
+ },
+ {
+ "entropy": 0.27788355611264703,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.8024702668190002,
+ "learning_rate": 1.8230943135928133e-05,
+ "loss": 0.21420254707336425,
+ "mean_token_accuracy": 0.9291928946971894,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.35886382102296593,
+ "eval_loss": 0.7313408851623535,
+ "eval_mean_token_accuracy": 0.8322128275137269,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 98.4564,
+ "eval_samples_per_second": 14.423,
+ "eval_steps_per_second": 1.808,
+ "step": 2900
+ },
+ {
+ "entropy": 0.24298458527296019,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.8567366600036621,
+ "learning_rate": 1.7705383764572093e-05,
+ "loss": 0.15583932399749756,
+ "mean_token_accuracy": 0.9481525069628006,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.3236005037520709,
+ "eval_loss": 0.8240236043930054,
+ "eval_mean_token_accuracy": 0.8270553162928378,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 98.3819,
+ "eval_samples_per_second": 14.434,
+ "eval_steps_per_second": 1.809,
+ "step": 2920
+ },
+ {
+ "entropy": 0.2195398263633251,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.8163788914680481,
+ "learning_rate": 1.7185042720186728e-05,
+ "loss": 0.14468494653701783,
+ "mean_token_accuracy": 0.9525706857442856,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.32686926424503326,
+ "eval_loss": 0.8029564023017883,
+ "eval_mean_token_accuracy": 0.8280681498934713,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 98.7239,
+ "eval_samples_per_second": 14.384,
+ "eval_steps_per_second": 1.803,
+ "step": 2940
+ },
+ {
+ "entropy": 0.21137732304632664,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.7098442316055298,
+ "learning_rate": 1.6670067253132565e-05,
+ "loss": 0.1382979393005371,
+ "mean_token_accuracy": 0.9557823866605759,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.32376345288887454,
+ "eval_loss": 0.8175393342971802,
+ "eval_mean_token_accuracy": 0.8281539598207796,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 98.4728,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 2960
+ },
+ {
+ "entropy": 0.21736963428556919,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.8196840286254883,
+ "learning_rate": 1.616060309537518e-05,
+ "loss": 0.14342793226242065,
+ "mean_token_accuracy": 0.9531146377325058,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.3240856607810835,
+ "eval_loss": 0.816273033618927,
+ "eval_mean_token_accuracy": 0.8283853041991759,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 98.4989,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 2980
+ },
+ {
+ "entropy": 0.2128997717052698,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.7581701278686523,
+ "learning_rate": 1.5656794419244872e-05,
+ "loss": 0.14148796796798707,
+ "mean_token_accuracy": 0.9551757454872132,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.3229605887043342,
+ "eval_loss": 0.8140764832496643,
+ "eval_mean_token_accuracy": 0.8288295952121864,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 98.4981,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 3000
+ },
+ {
+ "entropy": 0.214377661049366,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.6986305117607117,
+ "learning_rate": 1.5158783796637473e-05,
+ "loss": 0.14369987249374389,
+ "mean_token_accuracy": 0.9528444185853004,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.32001526101251665,
+ "eval_loss": 0.8190910816192627,
+ "eval_mean_token_accuracy": 0.829054100459881,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 98.5783,
+ "eval_samples_per_second": 14.405,
+ "eval_steps_per_second": 1.806,
+ "step": 3020
+ },
+ {
+ "entropy": 0.20811213366687298,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.9613916277885437,
+ "learning_rate": 1.4666712158668313e-05,
+ "loss": 0.14105329513549805,
+ "mean_token_accuracy": 0.9532626882195473,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.31941057892327895,
+ "eval_loss": 0.8219555616378784,
+ "eval_mean_token_accuracy": 0.8282468245270547,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 98.5199,
+ "eval_samples_per_second": 14.413,
+ "eval_steps_per_second": 1.807,
+ "step": 3040
+ },
+ {
+ "entropy": 0.2161025382578373,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.802244246006012,
+ "learning_rate": 1.418071875579028e-05,
+ "loss": 0.14670108556747435,
+ "mean_token_accuracy": 0.9525101512670517,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.3249504163526417,
+ "eval_loss": 0.8168519139289856,
+ "eval_mean_token_accuracy": 0.8281265336476015,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 98.4837,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 3060
+ },
+ {
+ "entropy": 0.21994274333119393,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.847614586353302,
+ "learning_rate": 1.370094111838762e-05,
+ "loss": 0.14627058506011964,
+ "mean_token_accuracy": 0.9523571625351905,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.3224444846423824,
+ "eval_loss": 0.8206745982170105,
+ "eval_mean_token_accuracy": 0.8285097687431936,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 98.5699,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 3080
+ },
+ {
+ "entropy": 0.21474989503622055,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.8615112900733948,
+ "learning_rate": 1.3227515017856523e-05,
+ "loss": 0.14166855812072754,
+ "mean_token_accuracy": 0.9545643255114555,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.3131480763634939,
+ "eval_loss": 0.8318397402763367,
+ "eval_mean_token_accuracy": 0.8293128348468395,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 98.5748,
+ "eval_samples_per_second": 14.405,
+ "eval_steps_per_second": 1.806,
+ "step": 3100
+ },
+ {
+ "entropy": 0.20732707269489764,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.8595603108406067,
+ "learning_rate": 1.2760574428183327e-05,
+ "loss": 0.1390355110168457,
+ "mean_token_accuracy": 0.9545612305402755,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.32303216425555475,
+ "eval_loss": 0.8206163048744202,
+ "eval_mean_token_accuracy": 0.8276272057817223,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 98.6296,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 3120
+ },
+ {
+ "entropy": 0.20272110737860202,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.9171860814094543,
+ "learning_rate": 1.2300251488031559e-05,
+ "loss": 0.13609397411346436,
+ "mean_token_accuracy": 0.9557161509990693,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.31589331360680334,
+ "eval_loss": 0.8248094916343689,
+ "eval_mean_token_accuracy": 0.8292221159077762,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 98.4574,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 3140
+ },
+ {
+ "entropy": 0.21678446158766745,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.870278000831604,
+ "learning_rate": 1.1846676463348161e-05,
+ "loss": 0.14677901268005372,
+ "mean_token_accuracy": 0.9534812569618225,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.3183298571391052,
+ "eval_loss": 0.8235951662063599,
+ "eval_mean_token_accuracy": 0.8291523583819357,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 98.5743,
+ "eval_samples_per_second": 14.405,
+ "eval_steps_per_second": 1.806,
+ "step": 3160
+ },
+ {
+ "entropy": 0.2208289708942175,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.8425227999687195,
+ "learning_rate": 1.1399977710499864e-05,
+ "loss": 0.146539568901062,
+ "mean_token_accuracy": 0.9524248003959656,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.3229933256346188,
+ "eval_loss": 0.8207331299781799,
+ "eval_mean_token_accuracy": 0.8284426217668512,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 98.6951,
+ "eval_samples_per_second": 14.388,
+ "eval_steps_per_second": 1.804,
+ "step": 3180
+ },
+ {
+ "entropy": 0.2222116433084011,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.8922525644302368,
+ "learning_rate": 1.0960281639949764e-05,
+ "loss": 0.14516948461532592,
+ "mean_token_accuracy": 0.952321757376194,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.32119952912410993,
+ "eval_loss": 0.8213328123092651,
+ "eval_mean_token_accuracy": 0.8285770188556628,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 98.6437,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 3200
+ },
+ {
+ "entropy": 0.21560499407351016,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.9485028386116028,
+ "learning_rate": 1.0527712680484752e-05,
+ "loss": 0.1465426802635193,
+ "mean_token_accuracy": 0.9517758965492249,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.3180267366298129,
+ "eval_loss": 0.8266745805740356,
+ "eval_mean_token_accuracy": 0.8283248446630628,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 98.5477,
+ "eval_samples_per_second": 14.409,
+ "eval_steps_per_second": 1.806,
+ "step": 3220
+ },
+ {
+ "entropy": 0.20934636779129506,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.7967023253440857,
+ "learning_rate": 1.0102393244003528e-05,
+ "loss": 0.14084668159484864,
+ "mean_token_accuracy": 0.9531109407544136,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.31779277040047593,
+ "eval_loss": 0.8262729048728943,
+ "eval_mean_token_accuracy": 0.8288095492995187,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 98.6036,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 3240
+ },
+ {
+ "entropy": 0.215614727884531,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.7962915301322937,
+ "learning_rate": 9.684443690875602e-06,
+ "loss": 0.14136850833892822,
+ "mean_token_accuracy": 0.9531293407082557,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.31656171481930806,
+ "eval_loss": 0.8303171992301941,
+ "eval_mean_token_accuracy": 0.8285069612974532,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 98.5105,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 3260
+ },
+ {
+ "entropy": 0.21908088140189647,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.9550889134407043,
+ "learning_rate": 9.273982295880618e-06,
+ "loss": 0.14732669591903685,
+ "mean_token_accuracy": 0.9522618219256401,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.3193875088283185,
+ "eval_loss": 0.8240101933479309,
+ "eval_mean_token_accuracy": 0.8285468552219734,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 98.4506,
+ "eval_samples_per_second": 14.423,
+ "eval_steps_per_second": 1.808,
+ "step": 3280
+ },
+ {
+ "entropy": 0.21472855806350707,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.8032136559486389,
+ "learning_rate": 8.871125214738089e-06,
+ "loss": 0.1441211223602295,
+ "mean_token_accuracy": 0.9532426640391349,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.3145652976766061,
+ "eval_loss": 0.8368120789527893,
+ "eval_mean_token_accuracy": 0.8285725685987579,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 98.5667,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 3300
+ },
+ {
+ "entropy": 0.20777193781657097,
+ "epoch": 8.0,
+ "grad_norm": 1.9440633058547974,
+ "learning_rate": 8.475986451236623e-06,
+ "loss": 0.13878930807113649,
+ "mean_token_accuracy": 0.9542667560088329,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.3138233663493328,
+ "eval_loss": 0.8311719298362732,
+ "eval_mean_token_accuracy": 0.8295681362741449,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 98.401,
+ "eval_samples_per_second": 14.431,
+ "eval_steps_per_second": 1.809,
+ "step": 3320
+ },
+ {
+ "entropy": 0.19024005569517613,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.6860044598579407,
+ "learning_rate": 8.088677824972262e-06,
+ "loss": 0.104938805103302,
+ "mean_token_accuracy": 0.9669251158833504,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.2925276593043563,
+ "eval_loss": 0.9170228838920593,
+ "eval_mean_token_accuracy": 0.8260884147681548,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 98.3738,
+ "eval_samples_per_second": 14.435,
+ "eval_steps_per_second": 1.809,
+ "step": 3340
+ },
+ {
+ "entropy": 0.17429662458598613,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.6853544116020203,
+ "learning_rate": 7.70930893970481e-06,
+ "loss": 0.09820220470428467,
+ "mean_token_accuracy": 0.9690332591533661,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.29154844699280985,
+ "eval_loss": 0.9182955026626587,
+ "eval_mean_token_accuracy": 0.8260592245653774,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 98.5024,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 3360
+ },
+ {
+ "entropy": 0.17998851463198662,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.6730113625526428,
+ "learning_rate": 7.337987152341293e-06,
+ "loss": 0.10509306192398071,
+ "mean_token_accuracy": 0.9663080736994744,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.29224191346530165,
+ "eval_loss": 0.9142967462539673,
+ "eval_mean_token_accuracy": 0.825743991672323,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 98.3875,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 3380
+ },
+ {
+ "entropy": 0.16425799671560526,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.8334925174713135,
+ "learning_rate": 6.974817542555239e-06,
+ "loss": 0.09438974857330322,
+ "mean_token_accuracy": 0.9704208940267562,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.2905296404542548,
+ "eval_loss": 0.9226241707801819,
+ "eval_mean_token_accuracy": 0.8262435773785195,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 98.6148,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 3400
+ },
+ {
+ "entropy": 0.16865960750728845,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.7445836663246155,
+ "learning_rate": 6.619902883050312e-06,
+ "loss": 0.09588083028793334,
+ "mean_token_accuracy": 0.9690534204244614,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.28748046489578954,
+ "eval_loss": 0.9263729453086853,
+ "eval_mean_token_accuracy": 0.8265480908115258,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 98.5098,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 3420
+ },
+ {
+ "entropy": 0.1850855764001608,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.8839195370674133,
+ "learning_rate": 6.273343610476938e-06,
+ "loss": 0.1067918062210083,
+ "mean_token_accuracy": 0.9668746247887612,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.2879818446850509,
+ "eval_loss": 0.9265933036804199,
+ "eval_mean_token_accuracy": 0.826676292700714,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 98.491,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 3440
+ },
+ {
+ "entropy": 0.1764491790905595,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.8340020775794983,
+ "learning_rate": 5.935237797009842e-06,
+ "loss": 0.10066965818405152,
+ "mean_token_accuracy": 0.9678841292858124,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.28914196303721224,
+ "eval_loss": 0.9251464605331421,
+ "eval_mean_token_accuracy": 0.8263801664448855,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 98.366,
+ "eval_samples_per_second": 14.436,
+ "eval_steps_per_second": 1.81,
+ "step": 3460
+ },
+ {
+ "entropy": 0.17226074449717999,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.7149388194084167,
+ "learning_rate": 5.605681122594823e-06,
+ "loss": 0.10052882432937622,
+ "mean_token_accuracy": 0.9679374754428863,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.28626472519689733,
+ "eval_loss": 0.9307431578636169,
+ "eval_mean_token_accuracy": 0.8269837619213576,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 98.3539,
+ "eval_samples_per_second": 14.438,
+ "eval_steps_per_second": 1.81,
+ "step": 3480
+ },
+ {
+ "entropy": 0.17341369036585091,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.6688793301582336,
+ "learning_rate": 5.28476684787248e-06,
+ "loss": 0.10054340362548828,
+ "mean_token_accuracy": 0.9675475805997849,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.28786867212378575,
+ "eval_loss": 0.9303860664367676,
+ "eval_mean_token_accuracy": 0.8259982466697693,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 98.627,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 3500
+ },
+ {
+ "entropy": 0.16994166951626538,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.7877247333526611,
+ "learning_rate": 4.972585787786484e-06,
+ "loss": 0.0993818998336792,
+ "mean_token_accuracy": 0.9688723921775818,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.287084077600013,
+ "eval_loss": 0.9293763041496277,
+ "eval_mean_token_accuracy": 0.8268557224380836,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 98.5488,
+ "eval_samples_per_second": 14.409,
+ "eval_steps_per_second": 1.806,
+ "step": 3520
+ },
+ {
+ "entropy": 0.1698782866820693,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.6359167695045471,
+ "learning_rate": 4.66922628588412e-06,
+ "loss": 0.10011144876480102,
+ "mean_token_accuracy": 0.968739403784275,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.2874377721481109,
+ "eval_loss": 0.9322494268417358,
+ "eval_mean_token_accuracy": 0.8263491124249576,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 98.3916,
+ "eval_samples_per_second": 14.432,
+ "eval_steps_per_second": 1.809,
+ "step": 3540
+ },
+ {
+ "entropy": 0.17174113132059574,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.7202840447425842,
+ "learning_rate": 4.374774189316017e-06,
+ "loss": 0.10309455394744874,
+ "mean_token_accuracy": 0.9690345630049706,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.286078037673168,
+ "eval_loss": 0.9365394115447998,
+ "eval_mean_token_accuracy": 0.8265787784972888,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 98.5151,
+ "eval_samples_per_second": 14.414,
+ "eval_steps_per_second": 1.807,
+ "step": 3560
+ },
+ {
+ "entropy": 0.1779118437319994,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.7755381464958191,
+ "learning_rate": 4.08931282454246e-06,
+ "loss": 0.10361287593841553,
+ "mean_token_accuracy": 0.9668692052364349,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.2901240846749102,
+ "eval_loss": 0.9282606244087219,
+ "eval_mean_token_accuracy": 0.8257483565405513,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 98.3953,
+ "eval_samples_per_second": 14.432,
+ "eval_steps_per_second": 1.809,
+ "step": 3580
+ },
+ {
+ "entropy": 0.17399974688887596,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.6068329215049744,
+ "learning_rate": 3.8129229737530146e-06,
+ "loss": 0.10165332555770874,
+ "mean_token_accuracy": 0.9683542385697365,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.28656756911384923,
+ "eval_loss": 0.93599933385849,
+ "eval_mean_token_accuracy": 0.8266928818788422,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 98.4991,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 3600
+ },
+ {
+ "entropy": 0.17109180577099323,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.671509325504303,
+ "learning_rate": 3.5456828520060676e-06,
+ "loss": 0.10388398170471191,
+ "mean_token_accuracy": 0.9666652023792267,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.28785753225007754,
+ "eval_loss": 0.9345396161079407,
+ "eval_mean_token_accuracy": 0.8258346751164855,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 98.5634,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 3620
+ },
+ {
+ "entropy": 0.17080388870090246,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.9064943194389343,
+ "learning_rate": 3.2876680850949342e-06,
+ "loss": 0.09955486059188842,
+ "mean_token_accuracy": 0.9688052341341973,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.2863340745313784,
+ "eval_loss": 0.9368643760681152,
+ "eval_mean_token_accuracy": 0.826385034604019,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 98.4956,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 3640
+ },
+ {
+ "entropy": 0.1739195568487048,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.645410418510437,
+ "learning_rate": 3.0389516881466266e-06,
+ "loss": 0.1038020372390747,
+ "mean_token_accuracy": 0.965588228404522,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.2869644621282481,
+ "eval_loss": 0.9337670803070068,
+ "eval_mean_token_accuracy": 0.8266236021947325,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 98.3869,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 3660
+ },
+ {
+ "entropy": 0.16863423213362694,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.6110771298408508,
+ "learning_rate": 2.7996040449594445e-06,
+ "loss": 0.10157880783081055,
+ "mean_token_accuracy": 0.967547157406807,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.28859044577968257,
+ "eval_loss": 0.9297890067100525,
+ "eval_mean_token_accuracy": 0.8264903181724335,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 98.5054,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 3680
+ },
+ {
+ "entropy": 0.1767999932169914,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.6543433666229248,
+ "learning_rate": 2.5696928880852117e-06,
+ "loss": 0.10324364900588989,
+ "mean_token_accuracy": 0.9673267215490341,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.2885892426364877,
+ "eval_loss": 0.928764820098877,
+ "eval_mean_token_accuracy": 0.8265983676642514,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 98.4914,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 3700
+ },
+ {
+ "entropy": 0.16943628452718257,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.8403868079185486,
+ "learning_rate": 2.3492832796617217e-06,
+ "loss": 0.09948049783706665,
+ "mean_token_accuracy": 0.967706835269928,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.2884856275795551,
+ "eval_loss": 0.933276355266571,
+ "eval_mean_token_accuracy": 0.8261616119507993,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 98.6259,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 3720
+ },
+ {
+ "entropy": 0.17359357613783616,
+ "epoch": 9.012062726176115,
+ "grad_norm": 0.37213996052742004,
+ "learning_rate": 2.138437593000977e-06,
+ "loss": 0.10089871883392335,
+ "mean_token_accuracy": 0.9696638874518566,
+ "num_tokens": 9122597.0,
+ "step": 3740
+ },
+ {
+ "epoch": 9.012062726176115,
+ "eval_entropy": 0.2875421148505104,
+ "eval_loss": 0.9357590079307556,
+ "eval_mean_token_accuracy": 0.8264994859025719,
+ "eval_num_tokens": 9122597.0,
+ "eval_runtime": 98.523,
+ "eval_samples_per_second": 14.413,
+ "eval_steps_per_second": 1.807,
+ "step": 3740
+ },
+ {
+ "entropy": 0.1642639346420765,
+ "epoch": 9.060313630880579,
+ "grad_norm": 0.5511618852615356,
+ "learning_rate": 1.937215494938237e-06,
+ "loss": 0.08524587750434875,
+ "mean_token_accuracy": 0.9754035726189614,
+ "num_tokens": 9171477.0,
+ "step": 3760
+ },
+ {
+ "epoch": 9.060313630880579,
+ "eval_entropy": 0.2820046850469675,
+ "eval_loss": 0.9588018655776978,
+ "eval_mean_token_accuracy": 0.8257443647036392,
+ "eval_num_tokens": 9171477.0,
+ "eval_runtime": 98.3992,
+ "eval_samples_per_second": 14.431,
+ "eval_steps_per_second": 1.809,
+ "step": 3760
+ },
+ {
+ "entropy": 0.1630103051662445,
+ "epoch": 9.108564535585042,
+ "grad_norm": 0.5125291347503662,
+ "learning_rate": 1.7456739289470143e-06,
+ "loss": 0.0895907461643219,
+ "mean_token_accuracy": 0.973100857436657,
+ "num_tokens": 9217543.0,
+ "step": 3780
+ },
+ {
+ "epoch": 9.108564535585042,
+ "eval_entropy": 0.2786683516388529,
+ "eval_loss": 0.9743365049362183,
+ "eval_mean_token_accuracy": 0.8253507131940863,
+ "eval_num_tokens": 9217543.0,
+ "eval_runtime": 98.3882,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 3780
+ },
+ {
+ "entropy": 0.15873534101992845,
+ "epoch": 9.156815440289506,
+ "grad_norm": 0.5697775483131409,
+ "learning_rate": 1.5638670990247788e-06,
+ "loss": 0.08552997708320617,
+ "mean_token_accuracy": 0.973672679066658,
+ "num_tokens": 9268724.0,
+ "step": 3800
+ },
+ {
+ "epoch": 9.156815440289506,
+ "eval_entropy": 0.2781428360704626,
+ "eval_loss": 0.9790021777153015,
+ "eval_mean_token_accuracy": 0.8251898780967413,
+ "eval_num_tokens": 9268724.0,
+ "eval_runtime": 98.5449,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 3800
+ },
+ {
+ "entropy": 0.1606229657307267,
+ "epoch": 9.205066344993968,
+ "grad_norm": 0.5715742707252502,
+ "learning_rate": 1.3918464543538225e-06,
+ "loss": 0.08984670042991638,
+ "mean_token_accuracy": 0.9729854613542557,
+ "num_tokens": 9316805.0,
+ "step": 3820
+ },
+ {
+ "epoch": 9.205066344993968,
+ "eval_entropy": 0.2777416408899125,
+ "eval_loss": 0.9808777570724487,
+ "eval_mean_token_accuracy": 0.8252255926641185,
+ "eval_num_tokens": 9316805.0,
+ "eval_runtime": 98.627,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 3820
+ },
+ {
+ "entropy": 0.15027269292622805,
+ "epoch": 9.253317249698432,
+ "grad_norm": 0.3359828293323517,
+ "learning_rate": 1.229660674741786e-06,
+ "loss": 0.08229721784591675,
+ "mean_token_accuracy": 0.9754723474383354,
+ "num_tokens": 9369940.0,
+ "step": 3840
+ },
+ {
+ "epoch": 9.253317249698432,
+ "eval_entropy": 0.2768587637483404,
+ "eval_loss": 0.9837455749511719,
+ "eval_mean_token_accuracy": 0.825493957889214,
+ "eval_num_tokens": 9369940.0,
+ "eval_runtime": 98.534,
+ "eval_samples_per_second": 14.411,
+ "eval_steps_per_second": 1.806,
+ "step": 3840
+ },
+ {
+ "entropy": 0.16767099779099226,
+ "epoch": 9.301568154402895,
+ "grad_norm": 0.5708875060081482,
+ "learning_rate": 1.077355656845822e-06,
+ "loss": 0.08888025283813476,
+ "mean_token_accuracy": 0.9715678468346596,
+ "num_tokens": 9416788.0,
+ "step": 3860
+ },
+ {
+ "epoch": 9.301568154402895,
+ "eval_entropy": 0.27612990283229377,
+ "eval_loss": 0.9873999357223511,
+ "eval_mean_token_accuracy": 0.8252144948150335,
+ "eval_num_tokens": 9416788.0,
+ "eval_runtime": 98.3889,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 3860
+ },
+ {
+ "entropy": 0.1549418892711401,
+ "epoch": 9.349819059107359,
+ "grad_norm": 0.7601456046104431,
+ "learning_rate": 9.349745011844029e-07,
+ "loss": 0.08411069512367249,
+ "mean_token_accuracy": 0.9741103485226631,
+ "num_tokens": 9467862.0,
+ "step": 3880
+ },
+ {
+ "epoch": 9.349819059107359,
+ "eval_entropy": 0.2757877716689967,
+ "eval_loss": 0.9892058968544006,
+ "eval_mean_token_accuracy": 0.8250612017143978,
+ "eval_num_tokens": 9467862.0,
+ "eval_runtime": 98.3869,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 3880
+ },
+ {
+ "entropy": 0.16111517809331416,
+ "epoch": 9.398069963811821,
+ "grad_norm": 0.6774276494979858,
+ "learning_rate": 8.025574999403695e-07,
+ "loss": 0.08765592575073242,
+ "mean_token_accuracy": 0.9734174907207489,
+ "num_tokens": 9514611.0,
+ "step": 3900
+ },
+ {
+ "epoch": 9.398069963811821,
+ "eval_entropy": 0.275664389049739,
+ "eval_loss": 0.989257276058197,
+ "eval_mean_token_accuracy": 0.8252896914321385,
+ "eval_num_tokens": 9514611.0,
+ "eval_runtime": 98.4649,
+ "eval_samples_per_second": 14.421,
+ "eval_steps_per_second": 1.808,
+ "step": 3900
+ },
+ {
+ "entropy": 0.15701593235135078,
+ "epoch": 9.446320868516285,
+ "grad_norm": 0.6130298376083374,
+ "learning_rate": 6.801421255587377e-07,
+ "loss": 0.0853091835975647,
+ "mean_token_accuracy": 0.9740088582038879,
+ "num_tokens": 9563437.0,
+ "step": 3920
+ },
+ {
+ "epoch": 9.446320868516285,
+ "eval_entropy": 0.27559053881114787,
+ "eval_loss": 0.9901270866394043,
+ "eval_mean_token_accuracy": 0.8251507653948966,
+ "eval_num_tokens": 9563437.0,
+ "eval_runtime": 98.5285,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 3920
+ },
+ {
+ "entropy": 0.15848954878747462,
+ "epoch": 9.494571773220748,
+ "grad_norm": 0.5928868651390076,
+ "learning_rate": 5.677630201424204e-07,
+ "loss": 0.08467000126838684,
+ "mean_token_accuracy": 0.9725634664297104,
+ "num_tokens": 9614034.0,
+ "step": 3940
+ },
+ {
+ "epoch": 9.494571773220748,
+ "eval_entropy": 0.2750965692521481,
+ "eval_loss": 0.9914345145225525,
+ "eval_mean_token_accuracy": 0.8252089954494091,
+ "eval_num_tokens": 9614034.0,
+ "eval_runtime": 98.5049,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 3940
+ },
+ {
+ "entropy": 0.15644967313855887,
+ "epoch": 9.54282267792521,
+ "grad_norm": 0.5718610882759094,
+ "learning_rate": 4.6545198564893463e-07,
+ "loss": 0.08564762473106384,
+ "mean_token_accuracy": 0.9730244860053062,
+ "num_tokens": 9661266.0,
+ "step": 3960
+ },
+ {
+ "epoch": 9.54282267792521,
+ "eval_entropy": 0.27522430820076654,
+ "eval_loss": 0.9908939003944397,
+ "eval_mean_token_accuracy": 0.8251052994406625,
+ "eval_num_tokens": 9661266.0,
+ "eval_runtime": 98.3637,
+ "eval_samples_per_second": 14.436,
+ "eval_steps_per_second": 1.81,
+ "step": 3960
+ },
+ {
+ "entropy": 0.15196608882397414,
+ "epoch": 9.591073582629674,
+ "grad_norm": 0.752661406993866,
+ "learning_rate": 3.732379748908161e-07,
+ "loss": 0.08072127103805542,
+ "mean_token_accuracy": 0.9749975100159645,
+ "num_tokens": 9710458.0,
+ "step": 3980
+ },
+ {
+ "epoch": 9.591073582629674,
+ "eval_entropy": 0.2750371985890892,
+ "eval_loss": 0.9913897514343262,
+ "eval_mean_token_accuracy": 0.8248539731073915,
+ "eval_num_tokens": 9710458.0,
+ "eval_runtime": 98.4959,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 3980
+ },
+ {
+ "entropy": 0.16616767533123494,
+ "epoch": 9.639324487334138,
+ "grad_norm": 0.7107775211334229,
+ "learning_rate": 2.911470833423274e-07,
+ "loss": 0.08988997340202332,
+ "mean_token_accuracy": 0.9714749112725258,
+ "num_tokens": 9755797.0,
+ "step": 4000
+ },
+ {
+ "epoch": 9.639324487334138,
+ "eval_entropy": 0.27523158457172053,
+ "eval_loss": 0.9913759827613831,
+ "eval_mean_token_accuracy": 0.8250464505693885,
+ "eval_num_tokens": 9755797.0,
+ "eval_runtime": 98.5067,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 4000
+ },
+ {
+ "entropy": 0.15760798901319503,
+ "epoch": 9.687575392038601,
+ "grad_norm": 0.5563387870788574,
+ "learning_rate": 2.1920254175472834e-07,
+ "loss": 0.0877057135105133,
+ "mean_token_accuracy": 0.9734948605298996,
+ "num_tokens": 9805552.0,
+ "step": 4020
+ },
+ {
+ "epoch": 9.687575392038601,
+ "eval_entropy": 0.27528667726208655,
+ "eval_loss": 0.9914706349372864,
+ "eval_mean_token_accuracy": 0.8250808802883277,
+ "eval_num_tokens": 9805552.0,
+ "eval_runtime": 98.475,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 4020
+ },
+ {
+ "entropy": 0.161653464846313,
+ "epoch": 9.735826296743063,
+ "grad_norm": 0.5896415114402771,
+ "learning_rate": 1.5742470958227596e-07,
+ "loss": 0.08775997757911683,
+ "mean_token_accuracy": 0.9718310445547104,
+ "num_tokens": 9852843.0,
+ "step": 4040
+ },
+ {
+ "epoch": 9.735826296743063,
+ "eval_entropy": 0.27531274182073184,
+ "eval_loss": 0.9915319085121155,
+ "eval_mean_token_accuracy": 0.825061980258213,
+ "eval_num_tokens": 9852843.0,
+ "eval_runtime": 98.3873,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 4040
+ },
+ {
+ "entropy": 0.16010624952614308,
+ "epoch": 9.784077201447527,
+ "grad_norm": 0.6022347211837769,
+ "learning_rate": 1.058310692207063e-07,
+ "loss": 0.08761411905288696,
+ "mean_token_accuracy": 0.9729624018073082,
+ "num_tokens": 9900151.0,
+ "step": 4060
+ },
+ {
+ "epoch": 9.784077201447527,
+ "eval_entropy": 0.27524732446737504,
+ "eval_loss": 0.991264820098877,
+ "eval_mean_token_accuracy": 0.8251047007153544,
+ "eval_num_tokens": 9900151.0,
+ "eval_runtime": 98.4834,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 4060
+ },
+ {
+ "entropy": 0.159021489135921,
+ "epoch": 9.83232810615199,
+ "grad_norm": 0.48451653122901917,
+ "learning_rate": 6.443622105994901e-08,
+ "loss": 0.08291574716567993,
+ "mean_token_accuracy": 0.9726693674921989,
+ "num_tokens": 9952926.0,
+ "step": 4080
+ },
+ {
+ "epoch": 9.83232810615199,
+ "eval_entropy": 0.27521907144717955,
+ "eval_loss": 0.9918096661567688,
+ "eval_mean_token_accuracy": 0.8250070894032382,
+ "eval_num_tokens": 9952926.0,
+ "eval_runtime": 98.3568,
+ "eval_samples_per_second": 14.437,
+ "eval_steps_per_second": 1.81,
+ "step": 4080
+ },
+ {
+ "entropy": 0.15175675600767136,
+ "epoch": 9.880579010856454,
+ "grad_norm": 0.5009680390357971,
+ "learning_rate": 3.325187935238231e-08,
+ "loss": 0.08214685916900635,
+ "mean_token_accuracy": 0.9752551406621933,
+ "num_tokens": 10003720.0,
+ "step": 4100
+ },
+ {
+ "epoch": 9.880579010856454,
+ "eval_entropy": 0.27522189488236826,
+ "eval_loss": 0.9918575882911682,
+ "eval_mean_token_accuracy": 0.8249149091458052,
+ "eval_num_tokens": 10003720.0,
+ "eval_runtime": 98.4802,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 4100
+ },
+ {
+ "entropy": 0.15018076542764902,
+ "epoch": 9.928829915560916,
+ "grad_norm": 0.6191889047622681,
+ "learning_rate": 1.2286868897828347e-08,
+ "loss": 0.08095031380653381,
+ "mean_token_accuracy": 0.9739995509386062,
+ "num_tokens": 10057590.0,
+ "step": 4120
+ },
+ {
+ "epoch": 9.928829915560916,
+ "eval_entropy": 0.27517637836464337,
+ "eval_loss": 0.991787850856781,
+ "eval_mean_token_accuracy": 0.8249709994605418,
+ "eval_num_tokens": 10057590.0,
+ "eval_runtime": 98.5572,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 4120
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 2.085753869939712e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4140/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4140/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4140/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4140/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4140/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4140/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4140/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4140/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4140/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4140/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4140/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4140/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4140/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4140/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..cb12487f370720c7c3c597520579d2a0c74c6bd3
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4140/trainer_state.json
@@ -0,0 +1,4381 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 9.97708082026538,
+ "eval_steps": 20,
+ "global_step": 4140,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ },
+ {
+ "entropy": 0.6141144648194313,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6467089056968689,
+ "learning_rate": 7.078840074092574e-05,
+ "loss": 0.5548617362976074,
+ "mean_token_accuracy": 0.8342376738786698,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.5951976082967908,
+ "eval_loss": 0.5850171446800232,
+ "eval_mean_token_accuracy": 0.8303041870004675,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 98.6601,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 760
+ },
+ {
+ "entropy": 0.5961867064237595,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.41280242800712585,
+ "learning_rate": 7.061001336495465e-05,
+ "loss": 0.5511267185211182,
+ "mean_token_accuracy": 0.8392465204000473,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6167017998320333,
+ "eval_loss": 0.5826791524887085,
+ "eval_mean_token_accuracy": 0.8309678050239434,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 98.6564,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 780
+ },
+ {
+ "entropy": 0.6025628983974457,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.4078947901725769,
+ "learning_rate": 7.042187293157881e-05,
+ "loss": 0.5552026271820069,
+ "mean_token_accuracy": 0.8359778299927711,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.6047008484601974,
+ "eval_loss": 0.5811123847961426,
+ "eval_mean_token_accuracy": 0.8320730804057603,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 98.6033,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 800
+ },
+ {
+ "entropy": 0.6041995719075203,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.543890118598938,
+ "learning_rate": 7.022403268231535e-05,
+ "loss": 0.554384422302246,
+ "mean_token_accuracy": 0.8380729153752327,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.6005127917179901,
+ "eval_loss": 0.5800131559371948,
+ "eval_mean_token_accuracy": 0.8313186547059691,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 98.4119,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 820
+ },
+ {
+ "entropy": 0.567442889396961,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.5015465021133423,
+ "learning_rate": 7.001654860361456e-05,
+ "loss": 0.5208251476287842,
+ "mean_token_accuracy": 0.8442784960453327,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.5785037809878253,
+ "eval_loss": 0.5865591168403625,
+ "eval_mean_token_accuracy": 0.8307993857378371,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 98.5061,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 840
+ },
+ {
+ "entropy": 0.5552012033760547,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.5975943207740784,
+ "learning_rate": 6.979947941101646e-05,
+ "loss": 0.5064489841461182,
+ "mean_token_accuracy": 0.8468396887183189,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5844205239515626,
+ "eval_loss": 0.585284948348999,
+ "eval_mean_token_accuracy": 0.8309797065311604,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 98.4982,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 860
+ },
+ {
+ "entropy": 0.5401638634502888,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.4194582998752594,
+ "learning_rate": 6.957288653253486e-05,
+ "loss": 0.49037814140319824,
+ "mean_token_accuracy": 0.8528496414422989,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.574507936835289,
+ "eval_loss": 0.5836747288703918,
+ "eval_mean_token_accuracy": 0.8325241643391298,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 98.5592,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 880
+ },
+ {
+ "entropy": 0.5606910541653634,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5395355820655823,
+ "learning_rate": 6.933683409127401e-05,
+ "loss": 0.5089571475982666,
+ "mean_token_accuracy": 0.8459660410881042,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5779740023479033,
+ "eval_loss": 0.5816043615341187,
+ "eval_mean_token_accuracy": 0.8326318036304431,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 98.4386,
+ "eval_samples_per_second": 14.425,
+ "eval_steps_per_second": 1.808,
+ "step": 900
+ },
+ {
+ "entropy": 0.5380903735756875,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.46588417887687683,
+ "learning_rate": 6.909138888728257e-05,
+ "loss": 0.4902662754058838,
+ "mean_token_accuracy": 0.8493925213813782,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.570154261220707,
+ "eval_loss": 0.5827520489692688,
+ "eval_mean_token_accuracy": 0.8321914167216654,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 98.5674,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 920
+ },
+ {
+ "entropy": 0.558555257320404,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.44576048851013184,
+ "learning_rate": 6.883662037864996e-05,
+ "loss": 0.5000699520111084,
+ "mean_token_accuracy": 0.8475695163011551,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5715417749761196,
+ "eval_loss": 0.5831062197685242,
+ "eval_mean_token_accuracy": 0.8327409899636601,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 98.568,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 940
+ },
+ {
+ "entropy": 0.5582522705197335,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.38966405391693115,
+ "learning_rate": 6.857260066185045e-05,
+ "loss": 0.5132341861724854,
+ "mean_token_accuracy": 0.8466864034533501,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5823918579334624,
+ "eval_loss": 0.5797588229179382,
+ "eval_mean_token_accuracy": 0.8326037566313583,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 98.6222,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 960
+ },
+ {
+ "entropy": 0.5557353265583516,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.42300811409950256,
+ "learning_rate": 6.829940445134089e-05,
+ "loss": 0.5035375118255615,
+ "mean_token_accuracy": 0.8479988545179367,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5542296482605881,
+ "eval_loss": 0.5800604820251465,
+ "eval_mean_token_accuracy": 0.8336404163515969,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 98.615,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 980
+ },
+ {
+ "entropy": 0.5492161035537719,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.47482752799987793,
+ "learning_rate": 6.801710905841724e-05,
+ "loss": 0.5032052040100098,
+ "mean_token_accuracy": 0.8488596931099892,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5755258423558781,
+ "eval_loss": 0.5764774680137634,
+ "eval_mean_token_accuracy": 0.8336571763070781,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 98.4765,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5439704649150372,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3656543493270874,
+ "learning_rate": 6.772579436933644e-05,
+ "loss": 0.49903340339660646,
+ "mean_token_accuracy": 0.8499931409955025,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5712662409530597,
+ "eval_loss": 0.5777108073234558,
+ "eval_mean_token_accuracy": 0.8336941352051296,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 98.678,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5448883637785912,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.3990268111228943,
+ "learning_rate": 6.742554282270959e-05,
+ "loss": 0.4990490436553955,
+ "mean_token_accuracy": 0.8488436549901962,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5816385483139017,
+ "eval_loss": 0.5749375820159912,
+ "eval_mean_token_accuracy": 0.833801462074344,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 98.6209,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5575303137302399,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.45015832781791687,
+ "learning_rate": 6.711643938617279e-05,
+ "loss": 0.5061063289642334,
+ "mean_token_accuracy": 0.8462719902396202,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5666238901655326,
+ "eval_loss": 0.574820876121521,
+ "eval_mean_token_accuracy": 0.8339746617869045,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 98.5332,
+ "eval_samples_per_second": 14.411,
+ "eval_steps_per_second": 1.806,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5477629877626896,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.43583670258522034,
+ "learning_rate": 6.679857153234218e-05,
+ "loss": 0.4998013496398926,
+ "mean_token_accuracy": 0.8465577483177185,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5809191281206152,
+ "eval_loss": 0.5737497210502625,
+ "eval_mean_token_accuracy": 0.8338909467284599,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 98.5093,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5548589795827865,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.44145169854164124,
+ "learning_rate": 6.647202921406043e-05,
+ "loss": 0.5035099029541016,
+ "mean_token_accuracy": 0.8463503196835518,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5626686118292005,
+ "eval_loss": 0.5707846879959106,
+ "eval_mean_token_accuracy": 0.8349729531936432,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 98.473,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5568635657429695,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.38482189178466797,
+ "learning_rate": 6.613690483894096e-05,
+ "loss": 0.5090720176696777,
+ "mean_token_accuracy": 0.8481037884950637,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5752792709998871,
+ "eval_loss": 0.5706034898757935,
+ "eval_mean_token_accuracy": 0.8350465498613507,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 98.6524,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5644363485276699,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.4353690445423126,
+ "learning_rate": 6.57932932432178e-05,
+ "loss": 0.5078121185302734,
+ "mean_token_accuracy": 0.846393559873104,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5670730096570561,
+ "eval_loss": 0.5699080228805542,
+ "eval_mean_token_accuracy": 0.8354134288396728,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 98.5697,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 1140
+ },
+ {
+ "entropy": 0.547992604970932,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.4126238524913788,
+ "learning_rate": 6.544129166490802e-05,
+ "loss": 0.4943515777587891,
+ "mean_token_accuracy": 0.8496095523238182,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5552005295673114,
+ "eval_loss": 0.5671029090881348,
+ "eval_mean_token_accuracy": 0.8364275894807965,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 98.5412,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 1160
+ },
+ {
+ "entropy": 0.561221782118082,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.47753170132637024,
+ "learning_rate": 6.508099971629446e-05,
+ "loss": 0.5094280242919922,
+ "mean_token_accuracy": 0.8462468951940536,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5532578974962234,
+ "eval_loss": 0.5714129209518433,
+ "eval_mean_token_accuracy": 0.8357450737712089,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 98.5358,
+ "eval_samples_per_second": 14.411,
+ "eval_steps_per_second": 1.806,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5400963060557842,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.38486021757125854,
+ "learning_rate": 6.471251935573679e-05,
+ "loss": 0.49443836212158204,
+ "mean_token_accuracy": 0.8484652638435364,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.5697383115465722,
+ "eval_loss": 0.5681865215301514,
+ "eval_mean_token_accuracy": 0.8350923453154189,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 98.4136,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5381059750914574,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.49116969108581543,
+ "learning_rate": 6.43359548588183e-05,
+ "loss": 0.4878206729888916,
+ "mean_token_accuracy": 0.8522022038698196,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5614734472853414,
+ "eval_loss": 0.5665943622589111,
+ "eval_mean_token_accuracy": 0.8360684296388304,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 98.5682,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5569418989121914,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.39628803730010986,
+ "learning_rate": 6.39514127888373e-05,
+ "loss": 0.5087827682495117,
+ "mean_token_accuracy": 0.8476181954145432,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5740904657358534,
+ "eval_loss": 0.5637108683586121,
+ "eval_mean_token_accuracy": 0.8366153638684348,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 98.4904,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 1240
+ },
+ {
+ "entropy": 0.5067426371268737,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.4593591094017029,
+ "learning_rate": 6.355900196665094e-05,
+ "loss": 0.4515994071960449,
+ "mean_token_accuracy": 0.8621168640943674,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5176989454566763,
+ "eval_loss": 0.5825926661491394,
+ "eval_mean_token_accuracy": 0.8353930775369152,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 98.5654,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 1260
+ },
+ {
+ "entropy": 0.4712781891226768,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.46717822551727295,
+ "learning_rate": 6.315883343988008e-05,
+ "loss": 0.4228676795959473,
+ "mean_token_accuracy": 0.8672123551368713,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.5229383730486538,
+ "eval_loss": 0.5875843167304993,
+ "eval_mean_token_accuracy": 0.8347370524084969,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 98.5459,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 1280
+ },
+ {
+ "entropy": 0.48235836550593375,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.5073364973068237,
+ "learning_rate": 6.275102045148425e-05,
+ "loss": 0.4365375518798828,
+ "mean_token_accuracy": 0.8644895792007447,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.5151654947339819,
+ "eval_loss": 0.5826129913330078,
+ "eval_mean_token_accuracy": 0.8359152366606037,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 98.3713,
+ "eval_samples_per_second": 14.435,
+ "eval_steps_per_second": 1.809,
+ "step": 1300
+ },
+ {
+ "entropy": 0.4756111137568951,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.4896017909049988,
+ "learning_rate": 6.233567840771526e-05,
+ "loss": 0.43177642822265627,
+ "mean_token_accuracy": 0.8645799145102501,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5227621813838401,
+ "eval_loss": 0.5796895027160645,
+ "eval_mean_token_accuracy": 0.8355028448479899,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 98.5538,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 1320
+ },
+ {
+ "entropy": 0.47541345208883284,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5064273476600647,
+ "learning_rate": 6.191292484545842e-05,
+ "loss": 0.4289750099182129,
+ "mean_token_accuracy": 0.8659383848309516,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.5286844237132019,
+ "eval_loss": 0.579947829246521,
+ "eval_mean_token_accuracy": 0.8353635648663125,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 98.589,
+ "eval_samples_per_second": 14.403,
+ "eval_steps_per_second": 1.805,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4916382558643818,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.5225116014480591,
+ "learning_rate": 6.148287939897118e-05,
+ "loss": 0.44352192878723146,
+ "mean_token_accuracy": 0.8630174115300179,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.5264859509267165,
+ "eval_loss": 0.5774893164634705,
+ "eval_mean_token_accuracy": 0.8363003827882617,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 98.5858,
+ "eval_samples_per_second": 14.404,
+ "eval_steps_per_second": 1.806,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4990046709775925,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.548435628414154,
+ "learning_rate": 6.104566376602795e-05,
+ "loss": 0.4463675498962402,
+ "mean_token_accuracy": 0.8630342438817025,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.522659856616781,
+ "eval_loss": 0.5790286660194397,
+ "eval_mean_token_accuracy": 0.8355394277010071,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 98.4707,
+ "eval_samples_per_second": 14.421,
+ "eval_steps_per_second": 1.808,
+ "step": 1380
+ },
+ {
+ "entropy": 0.48943497613072395,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6222821474075317,
+ "learning_rate": 6.060140167348109e-05,
+ "loss": 0.4356812000274658,
+ "mean_token_accuracy": 0.8644352272152901,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.5065617703654793,
+ "eval_loss": 0.5842992067337036,
+ "eval_mean_token_accuracy": 0.8358401662847968,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 98.4775,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 1400
+ },
+ {
+ "entropy": 0.47648413479328156,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.5207330584526062,
+ "learning_rate": 6.015021884224766e-05,
+ "loss": 0.42724952697753904,
+ "mean_token_accuracy": 0.8663465172052384,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.5232780023237292,
+ "eval_loss": 0.5769163966178894,
+ "eval_mean_token_accuracy": 0.8356800042511372,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 98.5005,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4767202906310558,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4695364534854889,
+ "learning_rate": 5.969224295173193e-05,
+ "loss": 0.42672252655029297,
+ "mean_token_accuracy": 0.8657647699117661,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.5146279760291067,
+ "eval_loss": 0.577904224395752,
+ "eval_mean_token_accuracy": 0.8365512393163831,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 98.5825,
+ "eval_samples_per_second": 14.404,
+ "eval_steps_per_second": 1.806,
+ "step": 1440
+ },
+ {
+ "entropy": 0.4855267658829689,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6333373785018921,
+ "learning_rate": 5.92276036036936e-05,
+ "loss": 0.4310962677001953,
+ "mean_token_accuracy": 0.864738130569458,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5260473481389913,
+ "eval_loss": 0.5770025253295898,
+ "eval_mean_token_accuracy": 0.8357346339172191,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 98.462,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 1460
+ },
+ {
+ "entropy": 0.4917050123214722,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.5833342671394348,
+ "learning_rate": 5.875643228557199e-05,
+ "loss": 0.44533047676086424,
+ "mean_token_accuracy": 0.8642907112836837,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.5278292167722509,
+ "eval_loss": 0.5773910880088806,
+ "eval_mean_token_accuracy": 0.83580608548743,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 98.4817,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 1480
+ },
+ {
+ "entropy": 0.5025837384164333,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.5484253168106079,
+ "learning_rate": 5.827886233327671e-05,
+ "loss": 0.44471564292907717,
+ "mean_token_accuracy": 0.8624475196003913,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.5118448720889145,
+ "eval_loss": 0.5786464810371399,
+ "eval_mean_token_accuracy": 0.8370849305994055,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 98.5208,
+ "eval_samples_per_second": 14.413,
+ "eval_steps_per_second": 1.807,
+ "step": 1500
+ },
+ {
+ "entropy": 0.494297581911087,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5475512146949768,
+ "learning_rate": 5.7795028893455165e-05,
+ "loss": 0.44118366241455076,
+ "mean_token_accuracy": 0.8630154177546501,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.5250470112548785,
+ "eval_loss": 0.5746602416038513,
+ "eval_mean_token_accuracy": 0.8370710711800651,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 98.54,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 1520
+ },
+ {
+ "entropy": 0.4840923048555851,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.39450952410697937,
+ "learning_rate": 5.730506888524761e-05,
+ "loss": 0.443526554107666,
+ "mean_token_accuracy": 0.8627525731921196,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5305671740448876,
+ "eval_loss": 0.5697008371353149,
+ "eval_mean_token_accuracy": 0.837059904015466,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 98.6288,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 1540
+ },
+ {
+ "entropy": 0.49613324701786043,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.47117286920547485,
+ "learning_rate": 5.680912096154071e-05,
+ "loss": 0.44313778877258303,
+ "mean_token_accuracy": 0.8614323407411575,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.522762700580479,
+ "eval_loss": 0.571703314781189,
+ "eval_mean_token_accuracy": 0.8381530970669864,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 98.6493,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4979200787842274,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.5275628566741943,
+ "learning_rate": 5.630732546973038e-05,
+ "loss": 0.4462272167205811,
+ "mean_token_accuracy": 0.8615201056003571,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.5232290047942922,
+ "eval_loss": 0.568954586982727,
+ "eval_mean_token_accuracy": 0.8380421128835571,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 98.484,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4957688055932522,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5806206464767456,
+ "learning_rate": 5.5799824412005126e-05,
+ "loss": 0.44518418312072755,
+ "mean_token_accuracy": 0.860895662009716,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.5238539929805177,
+ "eval_loss": 0.573081910610199,
+ "eval_mean_token_accuracy": 0.8369396288743179,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 98.5731,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 1600
+ },
+ {
+ "entropy": 0.5005358196794987,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.5114495754241943,
+ "learning_rate": 5.528676140516111e-05,
+ "loss": 0.45078649520874026,
+ "mean_token_accuracy": 0.860628567636013,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.5245393164707034,
+ "eval_loss": 0.5683825612068176,
+ "eval_mean_token_accuracy": 0.8380641599049728,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 98.5275,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 1620
+ },
+ {
+ "entropy": 0.4839292488992214,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6169933080673218,
+ "learning_rate": 5.4768281639960344e-05,
+ "loss": 0.43407382965087893,
+ "mean_token_accuracy": 0.867341585457325,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.5277225765284528,
+ "eval_loss": 0.567344069480896,
+ "eval_mean_token_accuracy": 0.8386145431673928,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 98.5702,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 1640
+ },
+ {
+ "entropy": 0.4952052327302786,
+ "epoch": 4.0,
+ "grad_norm": 2.080777168273926,
+ "learning_rate": 5.424453184004336e-05,
+ "loss": 0.4525449752807617,
+ "mean_token_accuracy": 0.8622837005517422,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.5197472232446242,
+ "eval_loss": 0.57077956199646,
+ "eval_mean_token_accuracy": 0.8381094560864266,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 98.5498,
+ "eval_samples_per_second": 14.409,
+ "eval_steps_per_second": 1.806,
+ "step": 1660
+ },
+ {
+ "entropy": 0.42416453957557676,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.41175922751426697,
+ "learning_rate": 5.3715660220408146e-05,
+ "loss": 0.3610882043838501,
+ "mean_token_accuracy": 0.8832842499017716,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.4814340365066957,
+ "eval_loss": 0.5963674783706665,
+ "eval_mean_token_accuracy": 0.8364584405770462,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 98.4464,
+ "eval_samples_per_second": 14.424,
+ "eval_steps_per_second": 1.808,
+ "step": 1680
+ },
+ {
+ "entropy": 0.4193890765309334,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.4773254990577698,
+ "learning_rate": 5.3181816445467126e-05,
+ "loss": 0.36023797988891604,
+ "mean_token_accuracy": 0.884614959359169,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.47795870448096417,
+ "eval_loss": 0.6029345393180847,
+ "eval_mean_token_accuracy": 0.8354013039824668,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 98.5784,
+ "eval_samples_per_second": 14.405,
+ "eval_steps_per_second": 1.806,
+ "step": 1700
+ },
+ {
+ "entropy": 0.41848619729280473,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.7293462157249451,
+ "learning_rate": 5.264315158669384e-05,
+ "loss": 0.36621780395507814,
+ "mean_token_accuracy": 0.882377204298973,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4683974619661824,
+ "eval_loss": 0.6010014414787292,
+ "eval_mean_token_accuracy": 0.8358984790491254,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 98.5406,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 1720
+ },
+ {
+ "entropy": 0.41377718672156333,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5061255693435669,
+ "learning_rate": 5.209981807987152e-05,
+ "loss": 0.36024787425994875,
+ "mean_token_accuracy": 0.8844522640109063,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.4666026450609893,
+ "eval_loss": 0.6050319075584412,
+ "eval_mean_token_accuracy": 0.8357557922266843,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 98.6056,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 1740
+ },
+ {
+ "entropy": 0.4185848042368889,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.49589282274246216,
+ "learning_rate": 5.155196968195561e-05,
+ "loss": 0.36044313907623293,
+ "mean_token_accuracy": 0.8863146558403969,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.4660757115382827,
+ "eval_loss": 0.6089843511581421,
+ "eval_mean_token_accuracy": 0.8352006492989786,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 98.5425,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 1760
+ },
+ {
+ "entropy": 0.4249664135277271,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.6652432680130005,
+ "learning_rate": 5.099976142756227e-05,
+ "loss": 0.37266385555267334,
+ "mean_token_accuracy": 0.880594964325428,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.48157174483443915,
+ "eval_loss": 0.5987246036529541,
+ "eval_mean_token_accuracy": 0.8361060595914219,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 98.4652,
+ "eval_samples_per_second": 14.421,
+ "eval_steps_per_second": 1.808,
+ "step": 1780
+ },
+ {
+ "entropy": 0.4195803053677082,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.6217629313468933,
+ "learning_rate": 5.044334958509551e-05,
+ "loss": 0.3646178483963013,
+ "mean_token_accuracy": 0.8808894723653793,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.47762840182593697,
+ "eval_loss": 0.5994009375572205,
+ "eval_mean_token_accuracy": 0.8366789422678144,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 98.6677,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 1800
+ },
+ {
+ "entropy": 0.42473565861582757,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.6135475635528564,
+ "learning_rate": 4.9882891612525023e-05,
+ "loss": 0.37249753475189207,
+ "mean_token_accuracy": 0.8800962314009666,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.47359335924802204,
+ "eval_loss": 0.6024664640426636,
+ "eval_mean_token_accuracy": 0.8355274846714534,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 98.5103,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 1820
+ },
+ {
+ "entropy": 0.4212189674377441,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.5874977707862854,
+ "learning_rate": 4.931854611282748e-05,
+ "loss": 0.36766781806945803,
+ "mean_token_accuracy": 0.8837772741913795,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.46670635798004234,
+ "eval_loss": 0.6085236668586731,
+ "eval_mean_token_accuracy": 0.8355892652206207,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 98.5871,
+ "eval_samples_per_second": 14.404,
+ "eval_steps_per_second": 1.806,
+ "step": 1840
+ },
+ {
+ "entropy": 0.41604730263352396,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5763360261917114,
+ "learning_rate": 4.875047278910371e-05,
+ "loss": 0.36643099784851074,
+ "mean_token_accuracy": 0.8825299561023712,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.47700903121005284,
+ "eval_loss": 0.606238842010498,
+ "eval_mean_token_accuracy": 0.8348092550642034,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 98.5305,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 1860
+ },
+ {
+ "entropy": 0.4189125746488571,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5916087627410889,
+ "learning_rate": 4.817883239938471e-05,
+ "loss": 0.3703711748123169,
+ "mean_token_accuracy": 0.8825044095516205,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.47982015703501324,
+ "eval_loss": 0.6003754138946533,
+ "eval_mean_token_accuracy": 0.8351969742373134,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 98.4111,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 1880
+ },
+ {
+ "entropy": 0.4429341182112694,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.6441562175750732,
+ "learning_rate": 4.7603786711138866e-05,
+ "loss": 0.3824815511703491,
+ "mean_token_accuracy": 0.8784194126725197,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.4740966230296017,
+ "eval_loss": 0.5988267660140991,
+ "eval_mean_token_accuracy": 0.8369577316085944,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 98.5291,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 1900
+ },
+ {
+ "entropy": 0.41137003526091576,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.6663628816604614,
+ "learning_rate": 4.702549845549385e-05,
+ "loss": 0.36178064346313477,
+ "mean_token_accuracy": 0.8830229461193084,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.4730259802904022,
+ "eval_loss": 0.6001402139663696,
+ "eval_mean_token_accuracy": 0.8366120139534554,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 98.5556,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 1920
+ },
+ {
+ "entropy": 0.4264552935957909,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6708512902259827,
+ "learning_rate": 4.64441312811855e-05,
+ "loss": 0.376303505897522,
+ "mean_token_accuracy": 0.8826978862285614,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4687648815720269,
+ "eval_loss": 0.5982367992401123,
+ "eval_mean_token_accuracy": 0.8368918728292658,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 98.4928,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 1940
+ },
+ {
+ "entropy": 0.4378764644265175,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.7098956108093262,
+ "learning_rate": 4.585984970824725e-05,
+ "loss": 0.38264145851135256,
+ "mean_token_accuracy": 0.8768767058849335,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.46603381248672354,
+ "eval_loss": 0.6010599732398987,
+ "eval_mean_token_accuracy": 0.8361037319965576,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 98.5157,
+ "eval_samples_per_second": 14.414,
+ "eval_steps_per_second": 1.807,
+ "step": 1960
+ },
+ {
+ "entropy": 0.44009693413972856,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.7529758214950562,
+ "learning_rate": 4.527281908145284e-05,
+ "loss": 0.379723048210144,
+ "mean_token_accuracy": 0.8799900904297828,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.47042709873633437,
+ "eval_loss": 0.5984888672828674,
+ "eval_mean_token_accuracy": 0.8375813492228475,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 98.3352,
+ "eval_samples_per_second": 14.44,
+ "eval_steps_per_second": 1.81,
+ "step": 1980
+ },
+ {
+ "entropy": 0.43963938802480695,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.530246376991272,
+ "learning_rate": 4.468320552352577e-05,
+ "loss": 0.3855365037918091,
+ "mean_token_accuracy": 0.8766040325164794,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4773505919435051,
+ "eval_loss": 0.5956445932388306,
+ "eval_mean_token_accuracy": 0.8365653522228926,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 98.6015,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 2000
+ },
+ {
+ "entropy": 0.4241503834724426,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.5264645218849182,
+ "learning_rate": 4.409117588812845e-05,
+ "loss": 0.3725905418395996,
+ "mean_token_accuracy": 0.8816388875246048,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4738255136468437,
+ "eval_loss": 0.5987666845321655,
+ "eval_mean_token_accuracy": 0.8362959995028678,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 98.6141,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 2020
+ },
+ {
+ "entropy": 0.4451489016413689,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.5322316884994507,
+ "learning_rate": 4.3496897712644684e-05,
+ "loss": 0.3903519630432129,
+ "mean_token_accuracy": 0.8752380400896073,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.4662560015247109,
+ "eval_loss": 0.5963740944862366,
+ "eval_mean_token_accuracy": 0.8378168083308788,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 98.6613,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 2040
+ },
+ {
+ "entropy": 0.43658976554870604,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.6989460587501526,
+ "learning_rate": 4.29005391707685e-05,
+ "loss": 0.3851702451705933,
+ "mean_token_accuracy": 0.8790383815765381,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4791082282749455,
+ "eval_loss": 0.5912667512893677,
+ "eval_mean_token_accuracy": 0.83739292286755,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 98.5606,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 2060
+ },
+ {
+ "entropy": 0.40598615316244274,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.5303810238838196,
+ "learning_rate": 4.2302269024913e-05,
+ "loss": 0.3450628280639648,
+ "mean_token_accuracy": 0.8893755567379487,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.42836171187711564,
+ "eval_loss": 0.6335396766662598,
+ "eval_mean_token_accuracy": 0.8349329397249757,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 98.3503,
+ "eval_samples_per_second": 14.438,
+ "eval_steps_per_second": 1.81,
+ "step": 2080
+ },
+ {
+ "entropy": 0.3517624609172344,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.6402326226234436,
+ "learning_rate": 4.170225657845261e-05,
+ "loss": 0.2898552894592285,
+ "mean_token_accuracy": 0.904600378870964,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.4094204810563098,
+ "eval_loss": 0.652080237865448,
+ "eval_mean_token_accuracy": 0.8331165705504042,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 98.3573,
+ "eval_samples_per_second": 14.437,
+ "eval_steps_per_second": 1.81,
+ "step": 2100
+ },
+ {
+ "entropy": 0.3560003057122231,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.9274800419807434,
+ "learning_rate": 4.110067162781221e-05,
+ "loss": 0.2926265954971313,
+ "mean_token_accuracy": 0.9040415585041046,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.41318756994906436,
+ "eval_loss": 0.6567395925521851,
+ "eval_mean_token_accuracy": 0.83362964197491,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 98.5042,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 2120
+ },
+ {
+ "entropy": 0.35080533251166346,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.6713427901268005,
+ "learning_rate": 4.049768441441677e-05,
+ "loss": 0.29822506904602053,
+ "mean_token_accuracy": 0.9027505293488503,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.416848015416874,
+ "eval_loss": 0.6563553214073181,
+ "eval_mean_token_accuracy": 0.8333374677079447,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 98.4261,
+ "eval_samples_per_second": 14.427,
+ "eval_steps_per_second": 1.808,
+ "step": 2140
+ },
+ {
+ "entropy": 0.3736354000866413,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.5667551755905151,
+ "learning_rate": 3.989346557651503e-05,
+ "loss": 0.3088909387588501,
+ "mean_token_accuracy": 0.9012129113078118,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.41974830309326727,
+ "eval_loss": 0.6504881381988525,
+ "eval_mean_token_accuracy": 0.8349264759695931,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 98.5007,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 2160
+ },
+ {
+ "entropy": 0.35063633099198344,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.7737690210342407,
+ "learning_rate": 3.9288186100890903e-05,
+ "loss": 0.2888066053390503,
+ "mean_token_accuracy": 0.9040960729122162,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.4058936465992017,
+ "eval_loss": 0.6578952670097351,
+ "eval_mean_token_accuracy": 0.8352396407823884,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 98.4807,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 2180
+ },
+ {
+ "entropy": 0.3545838728547096,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.8293929100036621,
+ "learning_rate": 3.8682017274476215e-05,
+ "loss": 0.29347708225250246,
+ "mean_token_accuracy": 0.902638104557991,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.41738414345832353,
+ "eval_loss": 0.6543260812759399,
+ "eval_mean_token_accuracy": 0.8336448260907376,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 98.499,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 2200
+ },
+ {
+ "entropy": 0.3416903533041477,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.7580088973045349,
+ "learning_rate": 3.8075130635878556e-05,
+ "loss": 0.28760058879852296,
+ "mean_token_accuracy": 0.906610831618309,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.4121073748958245,
+ "eval_loss": 0.6563934087753296,
+ "eval_mean_token_accuracy": 0.833708616120092,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 98.3808,
+ "eval_samples_per_second": 14.434,
+ "eval_steps_per_second": 1.809,
+ "step": 2220
+ },
+ {
+ "entropy": 0.352034630253911,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.7154259085655212,
+ "learning_rate": 3.746769792683787e-05,
+ "loss": 0.2967940092086792,
+ "mean_token_accuracy": 0.9023159235715866,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.4149914874454563,
+ "eval_loss": 0.6444063186645508,
+ "eval_mean_token_accuracy": 0.8360218017288809,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 98.4832,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 2240
+ },
+ {
+ "entropy": 0.3539439059793949,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.8078773617744446,
+ "learning_rate": 3.685989104362558e-05,
+ "loss": 0.2928986310958862,
+ "mean_token_accuracy": 0.905618867278099,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.4099769044793054,
+ "eval_loss": 0.6588519215583801,
+ "eval_mean_token_accuracy": 0.8353862765799748,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 98.3781,
+ "eval_samples_per_second": 14.434,
+ "eval_steps_per_second": 1.809,
+ "step": 2260
+ },
+ {
+ "entropy": 0.35821759328246117,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.8760809302330017,
+ "learning_rate": 3.6251881988399994e-05,
+ "loss": 0.29706525802612305,
+ "mean_token_accuracy": 0.9018172711133957,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.40791591667057425,
+ "eval_loss": 0.6583756804466248,
+ "eval_mean_token_accuracy": 0.8349186125096311,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 98.4808,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 2280
+ },
+ {
+ "entropy": 0.34672123566269875,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.8768810033798218,
+ "learning_rate": 3.564384282053165e-05,
+ "loss": 0.2885677576065063,
+ "mean_token_accuracy": 0.9063566774129868,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.412923522377282,
+ "eval_loss": 0.6535082459449768,
+ "eval_mean_token_accuracy": 0.8350685978873392,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 98.3842,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 2300
+ },
+ {
+ "entropy": 0.35606732666492463,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.7237412929534912,
+ "learning_rate": 3.503594560791267e-05,
+ "loss": 0.30418403148651124,
+ "mean_token_accuracy": 0.9004160806536674,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.42160589155856143,
+ "eval_loss": 0.6412338018417358,
+ "eval_mean_token_accuracy": 0.834980196162556,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 98.3674,
+ "eval_samples_per_second": 14.436,
+ "eval_steps_per_second": 1.81,
+ "step": 2320
+ },
+ {
+ "entropy": 0.35330103561282156,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.5864736437797546,
+ "learning_rate": 3.442836237826349e-05,
+ "loss": 0.29608047008514404,
+ "mean_token_accuracy": 0.9029312193393707,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.4273778943533308,
+ "eval_loss": 0.6407549977302551,
+ "eval_mean_token_accuracy": 0.8351332397943132,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 98.5533,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 2340
+ },
+ {
+ "entropy": 0.353700140863657,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.744305431842804,
+ "learning_rate": 3.382126507045109e-05,
+ "loss": 0.3004793643951416,
+ "mean_token_accuracy": 0.9015571728348732,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.4262272163053577,
+ "eval_loss": 0.6443405151367188,
+ "eval_mean_token_accuracy": 0.8354629797212193,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 98.3719,
+ "eval_samples_per_second": 14.435,
+ "eval_steps_per_second": 1.809,
+ "step": 2360
+ },
+ {
+ "entropy": 0.36056690439581873,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.6552110910415649,
+ "learning_rate": 3.321482548583228e-05,
+ "loss": 0.2950709104537964,
+ "mean_token_accuracy": 0.9033273816108703,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.41902332275771026,
+ "eval_loss": 0.6472867727279663,
+ "eval_mean_token_accuracy": 0.8354150207524889,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 98.3956,
+ "eval_samples_per_second": 14.432,
+ "eval_steps_per_second": 1.809,
+ "step": 2380
+ },
+ {
+ "entropy": 0.35097662582993505,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.670831024646759,
+ "learning_rate": 3.260921523963606e-05,
+ "loss": 0.29224560260772703,
+ "mean_token_accuracy": 0.9042141750454903,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.4284254325909561,
+ "eval_loss": 0.6407246589660645,
+ "eval_mean_token_accuracy": 0.8345498831084605,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 98.3651,
+ "eval_samples_per_second": 14.436,
+ "eval_steps_per_second": 1.81,
+ "step": 2400
+ },
+ {
+ "entropy": 0.35357567593455314,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.8287181258201599,
+ "learning_rate": 3.200460571239839e-05,
+ "loss": 0.2948127746582031,
+ "mean_token_accuracy": 0.9042218685150146,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.4094803942723221,
+ "eval_loss": 0.6497060060501099,
+ "eval_mean_token_accuracy": 0.8358801147910986,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 98.3681,
+ "eval_samples_per_second": 14.436,
+ "eval_steps_per_second": 1.81,
+ "step": 2420
+ },
+ {
+ "entropy": 0.36021770909428596,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.6800532937049866,
+ "learning_rate": 3.140116800146367e-05,
+ "loss": 0.30136709213256835,
+ "mean_token_accuracy": 0.9020729199051857,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.4202987858083811,
+ "eval_loss": 0.6428810954093933,
+ "eval_mean_token_accuracy": 0.8355459637856215,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 98.5147,
+ "eval_samples_per_second": 14.414,
+ "eval_steps_per_second": 1.807,
+ "step": 2440
+ },
+ {
+ "entropy": 0.34706664383411406,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.6911236643791199,
+ "learning_rate": 3.07990728725661e-05,
+ "loss": 0.28953113555908205,
+ "mean_token_accuracy": 0.9055185422301293,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.406153663825453,
+ "eval_loss": 0.6541930437088013,
+ "eval_mean_token_accuracy": 0.834371502814668,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 98.3494,
+ "eval_samples_per_second": 14.438,
+ "eval_steps_per_second": 1.81,
+ "step": 2460
+ },
+ {
+ "entropy": 0.34852043315768244,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.7544664144515991,
+ "learning_rate": 3.0198490711505038e-05,
+ "loss": 0.28950369358062744,
+ "mean_token_accuracy": 0.9057677388191223,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.4174931427735961,
+ "eval_loss": 0.6448875665664673,
+ "eval_mean_token_accuracy": 0.8358313615402478,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 98.4786,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 2480
+ },
+ {
+ "entropy": 0.31319292042499935,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.7616425156593323,
+ "learning_rate": 2.9599591475927893e-05,
+ "loss": 0.24788825511932372,
+ "mean_token_accuracy": 0.9198053983541635,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.3601247634780541,
+ "eval_loss": 0.7335909008979797,
+ "eval_mean_token_accuracy": 0.8309332149082356,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 98.3569,
+ "eval_samples_per_second": 14.437,
+ "eval_steps_per_second": 1.81,
+ "step": 2500
+ },
+ {
+ "entropy": 0.27905523739755156,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.7366685271263123,
+ "learning_rate": 2.9002544647234015e-05,
+ "loss": 0.21279211044311525,
+ "mean_token_accuracy": 0.929847776889801,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.36974305969275784,
+ "eval_loss": 0.7080109715461731,
+ "eval_mean_token_accuracy": 0.8330447111236915,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 98.4908,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 2520
+ },
+ {
+ "entropy": 0.2782863024622202,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.9167613387107849,
+ "learning_rate": 2.840751918261356e-05,
+ "loss": 0.20897085666656495,
+ "mean_token_accuracy": 0.9294204145669938,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.366466105319141,
+ "eval_loss": 0.7170661091804504,
+ "eval_mean_token_accuracy": 0.8328661680891273,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 98.4385,
+ "eval_samples_per_second": 14.425,
+ "eval_steps_per_second": 1.808,
+ "step": 2540
+ },
+ {
+ "entropy": 0.28355731777846815,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.8010057806968689,
+ "learning_rate": 2.7814683467234538e-05,
+ "loss": 0.211204195022583,
+ "mean_token_accuracy": 0.9297966167330742,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.36755596988656547,
+ "eval_loss": 0.7274795174598694,
+ "eval_mean_token_accuracy": 0.8303602230682802,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 98.487,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 2560
+ },
+ {
+ "entropy": 0.27822012901306153,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.7516391277313232,
+ "learning_rate": 2.7224205266591812e-05,
+ "loss": 0.2117089033126831,
+ "mean_token_accuracy": 0.9294167309999466,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.37035224410925016,
+ "eval_loss": 0.7123831510543823,
+ "eval_mean_token_accuracy": 0.8318134563022785,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 98.5029,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 2580
+ },
+ {
+ "entropy": 0.2784307342022657,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.8481637239456177,
+ "learning_rate": 2.6636251679031508e-05,
+ "loss": 0.21476459503173828,
+ "mean_token_accuracy": 0.9293668136000633,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.3675622491354353,
+ "eval_loss": 0.7221675515174866,
+ "eval_mean_token_accuracy": 0.8314910557832611,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 98.3892,
+ "eval_samples_per_second": 14.432,
+ "eval_steps_per_second": 1.809,
+ "step": 2600
+ },
+ {
+ "entropy": 0.27380565330386164,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.7563941478729248,
+ "learning_rate": 2.605098908846406e-05,
+ "loss": 0.2117525815963745,
+ "mean_token_accuracy": 0.9299218639731407,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.3676499978879864,
+ "eval_loss": 0.7249981164932251,
+ "eval_mean_token_accuracy": 0.8315788115678209,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 98.491,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 2620
+ },
+ {
+ "entropy": 0.26819119676947595,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.9161278605461121,
+ "learning_rate": 2.546858311727963e-05,
+ "loss": 0.2020425319671631,
+ "mean_token_accuracy": 0.9322459831833839,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.3589264298422953,
+ "eval_loss": 0.737293004989624,
+ "eval_mean_token_accuracy": 0.8307386482699534,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 98.4767,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 2640
+ },
+ {
+ "entropy": 0.275980044901371,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.7255368828773499,
+ "learning_rate": 2.488919857947885e-05,
+ "loss": 0.21207733154296876,
+ "mean_token_accuracy": 0.9280627354979515,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.3729423410102223,
+ "eval_loss": 0.7137707471847534,
+ "eval_mean_token_accuracy": 0.8315478007445175,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 98.384,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 2660
+ },
+ {
+ "entropy": 0.2794003043323755,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.817340612411499,
+ "learning_rate": 2.4312999434032465e-05,
+ "loss": 0.2155231714248657,
+ "mean_token_accuracy": 0.930096672475338,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.36857131709543506,
+ "eval_loss": 0.7273393869400024,
+ "eval_mean_token_accuracy": 0.8302122903673836,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 98.5119,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 2680
+ },
+ {
+ "entropy": 0.28550401143729687,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.7945013642311096,
+ "learning_rate": 2.374014873848286e-05,
+ "loss": 0.21428213119506836,
+ "mean_token_accuracy": 0.9295192703604698,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.3720237272843886,
+ "eval_loss": 0.7190349698066711,
+ "eval_mean_token_accuracy": 0.8307539365264807,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 98.3972,
+ "eval_samples_per_second": 14.431,
+ "eval_steps_per_second": 1.809,
+ "step": 2700
+ },
+ {
+ "entropy": 0.2789166793227196,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.7820216417312622,
+ "learning_rate": 2.3170808602800694e-05,
+ "loss": 0.21065421104431153,
+ "mean_token_accuracy": 0.9293019056320191,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.36198576715555086,
+ "eval_loss": 0.7250529527664185,
+ "eval_mean_token_accuracy": 0.8310985079642093,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 98.516,
+ "eval_samples_per_second": 14.414,
+ "eval_steps_per_second": 1.807,
+ "step": 2720
+ },
+ {
+ "entropy": 0.2805891681462526,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.8589067459106445,
+ "learning_rate": 2.2605140143509734e-05,
+ "loss": 0.21532771587371827,
+ "mean_token_accuracy": 0.9301707550883294,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.36839111149311066,
+ "eval_loss": 0.720038890838623,
+ "eval_mean_token_accuracy": 0.8314180384191234,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 98.3864,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 2740
+ },
+ {
+ "entropy": 0.2923452850431204,
+ "epoch": 6.651387213510254,
+ "grad_norm": 1.1464570760726929,
+ "learning_rate": 2.2043303438092746e-05,
+ "loss": 0.22156639099121095,
+ "mean_token_accuracy": 0.9259255439043045,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.36308020152402726,
+ "eval_loss": 0.7283808588981628,
+ "eval_mean_token_accuracy": 0.8307237183110098,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 98.3733,
+ "eval_samples_per_second": 14.435,
+ "eval_steps_per_second": 1.809,
+ "step": 2760
+ },
+ {
+ "entropy": 0.2749974336475134,
+ "epoch": 6.699638118214716,
+ "grad_norm": 1.0338026285171509,
+ "learning_rate": 2.1485457479691453e-05,
+ "loss": 0.2111821413040161,
+ "mean_token_accuracy": 0.9308380514383316,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.35679193979568696,
+ "eval_loss": 0.7316031455993652,
+ "eval_mean_token_accuracy": 0.8316076066386834,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 98.4012,
+ "eval_samples_per_second": 14.431,
+ "eval_steps_per_second": 1.809,
+ "step": 2780
+ },
+ {
+ "entropy": 0.2827217370271683,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.6860962510108948,
+ "learning_rate": 2.0931760132113435e-05,
+ "loss": 0.219130802154541,
+ "mean_token_accuracy": 0.9272046208381652,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.367448085479522,
+ "eval_loss": 0.7210394144058228,
+ "eval_mean_token_accuracy": 0.8324049028118005,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 98.3867,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 2800
+ },
+ {
+ "entropy": 0.28139964006841184,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.7198498249053955,
+ "learning_rate": 2.038236808515841e-05,
+ "loss": 0.2179172992706299,
+ "mean_token_accuracy": 0.9282332047820091,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.3658287344353922,
+ "eval_loss": 0.7204482555389404,
+ "eval_mean_token_accuracy": 0.8324238242727987,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 98.3796,
+ "eval_samples_per_second": 14.434,
+ "eval_steps_per_second": 1.809,
+ "step": 2820
+ },
+ {
+ "entropy": 0.28035391345620153,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.8470256924629211,
+ "learning_rate": 1.9837436810277026e-05,
+ "loss": 0.21534478664398193,
+ "mean_token_accuracy": 0.9284071952104569,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.3615414013018769,
+ "eval_loss": 0.7223248481750488,
+ "eval_mean_token_accuracy": 0.8328172581249409,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 98.4426,
+ "eval_samples_per_second": 14.425,
+ "eval_steps_per_second": 1.808,
+ "step": 2840
+ },
+ {
+ "entropy": 0.27447608187794686,
+ "epoch": 6.892641737032569,
+ "grad_norm": 1.1274205446243286,
+ "learning_rate": 1.9297120516574132e-05,
+ "loss": 0.2151768684387207,
+ "mean_token_accuracy": 0.9284873619675637,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.36038200892089456,
+ "eval_loss": 0.7199941873550415,
+ "eval_mean_token_accuracy": 0.8334562922461649,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 98.4807,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 2860
+ },
+ {
+ "entropy": 0.27447244003415106,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.857891857624054,
+ "learning_rate": 1.876157210716952e-05,
+ "loss": 0.21279585361480713,
+ "mean_token_accuracy": 0.9296008720993996,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.3578391455197602,
+ "eval_loss": 0.7355349063873291,
+ "eval_mean_token_accuracy": 0.8319483108065101,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 98.3412,
+ "eval_samples_per_second": 14.44,
+ "eval_steps_per_second": 1.81,
+ "step": 2880
+ },
+ {
+ "entropy": 0.27788355611264703,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.8024702668190002,
+ "learning_rate": 1.8230943135928133e-05,
+ "loss": 0.21420254707336425,
+ "mean_token_accuracy": 0.9291928946971894,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.35886382102296593,
+ "eval_loss": 0.7313408851623535,
+ "eval_mean_token_accuracy": 0.8322128275137269,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 98.4564,
+ "eval_samples_per_second": 14.423,
+ "eval_steps_per_second": 1.808,
+ "step": 2900
+ },
+ {
+ "entropy": 0.24298458527296019,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.8567366600036621,
+ "learning_rate": 1.7705383764572093e-05,
+ "loss": 0.15583932399749756,
+ "mean_token_accuracy": 0.9481525069628006,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.3236005037520709,
+ "eval_loss": 0.8240236043930054,
+ "eval_mean_token_accuracy": 0.8270553162928378,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 98.3819,
+ "eval_samples_per_second": 14.434,
+ "eval_steps_per_second": 1.809,
+ "step": 2920
+ },
+ {
+ "entropy": 0.2195398263633251,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.8163788914680481,
+ "learning_rate": 1.7185042720186728e-05,
+ "loss": 0.14468494653701783,
+ "mean_token_accuracy": 0.9525706857442856,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.32686926424503326,
+ "eval_loss": 0.8029564023017883,
+ "eval_mean_token_accuracy": 0.8280681498934713,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 98.7239,
+ "eval_samples_per_second": 14.384,
+ "eval_steps_per_second": 1.803,
+ "step": 2940
+ },
+ {
+ "entropy": 0.21137732304632664,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.7098442316055298,
+ "learning_rate": 1.6670067253132565e-05,
+ "loss": 0.1382979393005371,
+ "mean_token_accuracy": 0.9557823866605759,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.32376345288887454,
+ "eval_loss": 0.8175393342971802,
+ "eval_mean_token_accuracy": 0.8281539598207796,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 98.4728,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 2960
+ },
+ {
+ "entropy": 0.21736963428556919,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.8196840286254883,
+ "learning_rate": 1.616060309537518e-05,
+ "loss": 0.14342793226242065,
+ "mean_token_accuracy": 0.9531146377325058,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.3240856607810835,
+ "eval_loss": 0.816273033618927,
+ "eval_mean_token_accuracy": 0.8283853041991759,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 98.4989,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 2980
+ },
+ {
+ "entropy": 0.2128997717052698,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.7581701278686523,
+ "learning_rate": 1.5656794419244872e-05,
+ "loss": 0.14148796796798707,
+ "mean_token_accuracy": 0.9551757454872132,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.3229605887043342,
+ "eval_loss": 0.8140764832496643,
+ "eval_mean_token_accuracy": 0.8288295952121864,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 98.4981,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 3000
+ },
+ {
+ "entropy": 0.214377661049366,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.6986305117607117,
+ "learning_rate": 1.5158783796637473e-05,
+ "loss": 0.14369987249374389,
+ "mean_token_accuracy": 0.9528444185853004,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.32001526101251665,
+ "eval_loss": 0.8190910816192627,
+ "eval_mean_token_accuracy": 0.829054100459881,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 98.5783,
+ "eval_samples_per_second": 14.405,
+ "eval_steps_per_second": 1.806,
+ "step": 3020
+ },
+ {
+ "entropy": 0.20811213366687298,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.9613916277885437,
+ "learning_rate": 1.4666712158668313e-05,
+ "loss": 0.14105329513549805,
+ "mean_token_accuracy": 0.9532626882195473,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.31941057892327895,
+ "eval_loss": 0.8219555616378784,
+ "eval_mean_token_accuracy": 0.8282468245270547,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 98.5199,
+ "eval_samples_per_second": 14.413,
+ "eval_steps_per_second": 1.807,
+ "step": 3040
+ },
+ {
+ "entropy": 0.2161025382578373,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.802244246006012,
+ "learning_rate": 1.418071875579028e-05,
+ "loss": 0.14670108556747435,
+ "mean_token_accuracy": 0.9525101512670517,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.3249504163526417,
+ "eval_loss": 0.8168519139289856,
+ "eval_mean_token_accuracy": 0.8281265336476015,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 98.4837,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 3060
+ },
+ {
+ "entropy": 0.21994274333119393,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.847614586353302,
+ "learning_rate": 1.370094111838762e-05,
+ "loss": 0.14627058506011964,
+ "mean_token_accuracy": 0.9523571625351905,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.3224444846423824,
+ "eval_loss": 0.8206745982170105,
+ "eval_mean_token_accuracy": 0.8285097687431936,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 98.5699,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 3080
+ },
+ {
+ "entropy": 0.21474989503622055,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.8615112900733948,
+ "learning_rate": 1.3227515017856523e-05,
+ "loss": 0.14166855812072754,
+ "mean_token_accuracy": 0.9545643255114555,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.3131480763634939,
+ "eval_loss": 0.8318397402763367,
+ "eval_mean_token_accuracy": 0.8293128348468395,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 98.5748,
+ "eval_samples_per_second": 14.405,
+ "eval_steps_per_second": 1.806,
+ "step": 3100
+ },
+ {
+ "entropy": 0.20732707269489764,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.8595603108406067,
+ "learning_rate": 1.2760574428183327e-05,
+ "loss": 0.1390355110168457,
+ "mean_token_accuracy": 0.9545612305402755,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.32303216425555475,
+ "eval_loss": 0.8206163048744202,
+ "eval_mean_token_accuracy": 0.8276272057817223,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 98.6296,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 3120
+ },
+ {
+ "entropy": 0.20272110737860202,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.9171860814094543,
+ "learning_rate": 1.2300251488031559e-05,
+ "loss": 0.13609397411346436,
+ "mean_token_accuracy": 0.9557161509990693,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.31589331360680334,
+ "eval_loss": 0.8248094916343689,
+ "eval_mean_token_accuracy": 0.8292221159077762,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 98.4574,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 3140
+ },
+ {
+ "entropy": 0.21678446158766745,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.870278000831604,
+ "learning_rate": 1.1846676463348161e-05,
+ "loss": 0.14677901268005372,
+ "mean_token_accuracy": 0.9534812569618225,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.3183298571391052,
+ "eval_loss": 0.8235951662063599,
+ "eval_mean_token_accuracy": 0.8291523583819357,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 98.5743,
+ "eval_samples_per_second": 14.405,
+ "eval_steps_per_second": 1.806,
+ "step": 3160
+ },
+ {
+ "entropy": 0.2208289708942175,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.8425227999687195,
+ "learning_rate": 1.1399977710499864e-05,
+ "loss": 0.146539568901062,
+ "mean_token_accuracy": 0.9524248003959656,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.3229933256346188,
+ "eval_loss": 0.8207331299781799,
+ "eval_mean_token_accuracy": 0.8284426217668512,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 98.6951,
+ "eval_samples_per_second": 14.388,
+ "eval_steps_per_second": 1.804,
+ "step": 3180
+ },
+ {
+ "entropy": 0.2222116433084011,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.8922525644302368,
+ "learning_rate": 1.0960281639949764e-05,
+ "loss": 0.14516948461532592,
+ "mean_token_accuracy": 0.952321757376194,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.32119952912410993,
+ "eval_loss": 0.8213328123092651,
+ "eval_mean_token_accuracy": 0.8285770188556628,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 98.6437,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 3200
+ },
+ {
+ "entropy": 0.21560499407351016,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.9485028386116028,
+ "learning_rate": 1.0527712680484752e-05,
+ "loss": 0.1465426802635193,
+ "mean_token_accuracy": 0.9517758965492249,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.3180267366298129,
+ "eval_loss": 0.8266745805740356,
+ "eval_mean_token_accuracy": 0.8283248446630628,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 98.5477,
+ "eval_samples_per_second": 14.409,
+ "eval_steps_per_second": 1.806,
+ "step": 3220
+ },
+ {
+ "entropy": 0.20934636779129506,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.7967023253440857,
+ "learning_rate": 1.0102393244003528e-05,
+ "loss": 0.14084668159484864,
+ "mean_token_accuracy": 0.9531109407544136,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.31779277040047593,
+ "eval_loss": 0.8262729048728943,
+ "eval_mean_token_accuracy": 0.8288095492995187,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 98.6036,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 3240
+ },
+ {
+ "entropy": 0.215614727884531,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.7962915301322937,
+ "learning_rate": 9.684443690875602e-06,
+ "loss": 0.14136850833892822,
+ "mean_token_accuracy": 0.9531293407082557,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.31656171481930806,
+ "eval_loss": 0.8303171992301941,
+ "eval_mean_token_accuracy": 0.8285069612974532,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 98.5105,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 3260
+ },
+ {
+ "entropy": 0.21908088140189647,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.9550889134407043,
+ "learning_rate": 9.273982295880618e-06,
+ "loss": 0.14732669591903685,
+ "mean_token_accuracy": 0.9522618219256401,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.3193875088283185,
+ "eval_loss": 0.8240101933479309,
+ "eval_mean_token_accuracy": 0.8285468552219734,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 98.4506,
+ "eval_samples_per_second": 14.423,
+ "eval_steps_per_second": 1.808,
+ "step": 3280
+ },
+ {
+ "entropy": 0.21472855806350707,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.8032136559486389,
+ "learning_rate": 8.871125214738089e-06,
+ "loss": 0.1441211223602295,
+ "mean_token_accuracy": 0.9532426640391349,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.3145652976766061,
+ "eval_loss": 0.8368120789527893,
+ "eval_mean_token_accuracy": 0.8285725685987579,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 98.5667,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 3300
+ },
+ {
+ "entropy": 0.20777193781657097,
+ "epoch": 8.0,
+ "grad_norm": 1.9440633058547974,
+ "learning_rate": 8.475986451236623e-06,
+ "loss": 0.13878930807113649,
+ "mean_token_accuracy": 0.9542667560088329,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.3138233663493328,
+ "eval_loss": 0.8311719298362732,
+ "eval_mean_token_accuracy": 0.8295681362741449,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 98.401,
+ "eval_samples_per_second": 14.431,
+ "eval_steps_per_second": 1.809,
+ "step": 3320
+ },
+ {
+ "entropy": 0.19024005569517613,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.6860044598579407,
+ "learning_rate": 8.088677824972262e-06,
+ "loss": 0.104938805103302,
+ "mean_token_accuracy": 0.9669251158833504,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.2925276593043563,
+ "eval_loss": 0.9170228838920593,
+ "eval_mean_token_accuracy": 0.8260884147681548,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 98.3738,
+ "eval_samples_per_second": 14.435,
+ "eval_steps_per_second": 1.809,
+ "step": 3340
+ },
+ {
+ "entropy": 0.17429662458598613,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.6853544116020203,
+ "learning_rate": 7.70930893970481e-06,
+ "loss": 0.09820220470428467,
+ "mean_token_accuracy": 0.9690332591533661,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.29154844699280985,
+ "eval_loss": 0.9182955026626587,
+ "eval_mean_token_accuracy": 0.8260592245653774,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 98.5024,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 3360
+ },
+ {
+ "entropy": 0.17998851463198662,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.6730113625526428,
+ "learning_rate": 7.337987152341293e-06,
+ "loss": 0.10509306192398071,
+ "mean_token_accuracy": 0.9663080736994744,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.29224191346530165,
+ "eval_loss": 0.9142967462539673,
+ "eval_mean_token_accuracy": 0.825743991672323,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 98.3875,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 3380
+ },
+ {
+ "entropy": 0.16425799671560526,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.8334925174713135,
+ "learning_rate": 6.974817542555239e-06,
+ "loss": 0.09438974857330322,
+ "mean_token_accuracy": 0.9704208940267562,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.2905296404542548,
+ "eval_loss": 0.9226241707801819,
+ "eval_mean_token_accuracy": 0.8262435773785195,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 98.6148,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 3400
+ },
+ {
+ "entropy": 0.16865960750728845,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.7445836663246155,
+ "learning_rate": 6.619902883050312e-06,
+ "loss": 0.09588083028793334,
+ "mean_token_accuracy": 0.9690534204244614,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.28748046489578954,
+ "eval_loss": 0.9263729453086853,
+ "eval_mean_token_accuracy": 0.8265480908115258,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 98.5098,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 3420
+ },
+ {
+ "entropy": 0.1850855764001608,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.8839195370674133,
+ "learning_rate": 6.273343610476938e-06,
+ "loss": 0.1067918062210083,
+ "mean_token_accuracy": 0.9668746247887612,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.2879818446850509,
+ "eval_loss": 0.9265933036804199,
+ "eval_mean_token_accuracy": 0.826676292700714,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 98.491,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 3440
+ },
+ {
+ "entropy": 0.1764491790905595,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.8340020775794983,
+ "learning_rate": 5.935237797009842e-06,
+ "loss": 0.10066965818405152,
+ "mean_token_accuracy": 0.9678841292858124,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.28914196303721224,
+ "eval_loss": 0.9251464605331421,
+ "eval_mean_token_accuracy": 0.8263801664448855,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 98.366,
+ "eval_samples_per_second": 14.436,
+ "eval_steps_per_second": 1.81,
+ "step": 3460
+ },
+ {
+ "entropy": 0.17226074449717999,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.7149388194084167,
+ "learning_rate": 5.605681122594823e-06,
+ "loss": 0.10052882432937622,
+ "mean_token_accuracy": 0.9679374754428863,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.28626472519689733,
+ "eval_loss": 0.9307431578636169,
+ "eval_mean_token_accuracy": 0.8269837619213576,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 98.3539,
+ "eval_samples_per_second": 14.438,
+ "eval_steps_per_second": 1.81,
+ "step": 3480
+ },
+ {
+ "entropy": 0.17341369036585091,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.6688793301582336,
+ "learning_rate": 5.28476684787248e-06,
+ "loss": 0.10054340362548828,
+ "mean_token_accuracy": 0.9675475805997849,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.28786867212378575,
+ "eval_loss": 0.9303860664367676,
+ "eval_mean_token_accuracy": 0.8259982466697693,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 98.627,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 3500
+ },
+ {
+ "entropy": 0.16994166951626538,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.7877247333526611,
+ "learning_rate": 4.972585787786484e-06,
+ "loss": 0.0993818998336792,
+ "mean_token_accuracy": 0.9688723921775818,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.287084077600013,
+ "eval_loss": 0.9293763041496277,
+ "eval_mean_token_accuracy": 0.8268557224380836,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 98.5488,
+ "eval_samples_per_second": 14.409,
+ "eval_steps_per_second": 1.806,
+ "step": 3520
+ },
+ {
+ "entropy": 0.1698782866820693,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.6359167695045471,
+ "learning_rate": 4.66922628588412e-06,
+ "loss": 0.10011144876480102,
+ "mean_token_accuracy": 0.968739403784275,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.2874377721481109,
+ "eval_loss": 0.9322494268417358,
+ "eval_mean_token_accuracy": 0.8263491124249576,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 98.3916,
+ "eval_samples_per_second": 14.432,
+ "eval_steps_per_second": 1.809,
+ "step": 3540
+ },
+ {
+ "entropy": 0.17174113132059574,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.7202840447425842,
+ "learning_rate": 4.374774189316017e-06,
+ "loss": 0.10309455394744874,
+ "mean_token_accuracy": 0.9690345630049706,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.286078037673168,
+ "eval_loss": 0.9365394115447998,
+ "eval_mean_token_accuracy": 0.8265787784972888,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 98.5151,
+ "eval_samples_per_second": 14.414,
+ "eval_steps_per_second": 1.807,
+ "step": 3560
+ },
+ {
+ "entropy": 0.1779118437319994,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.7755381464958191,
+ "learning_rate": 4.08931282454246e-06,
+ "loss": 0.10361287593841553,
+ "mean_token_accuracy": 0.9668692052364349,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.2901240846749102,
+ "eval_loss": 0.9282606244087219,
+ "eval_mean_token_accuracy": 0.8257483565405513,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 98.3953,
+ "eval_samples_per_second": 14.432,
+ "eval_steps_per_second": 1.809,
+ "step": 3580
+ },
+ {
+ "entropy": 0.17399974688887596,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.6068329215049744,
+ "learning_rate": 3.8129229737530146e-06,
+ "loss": 0.10165332555770874,
+ "mean_token_accuracy": 0.9683542385697365,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.28656756911384923,
+ "eval_loss": 0.93599933385849,
+ "eval_mean_token_accuracy": 0.8266928818788422,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 98.4991,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 3600
+ },
+ {
+ "entropy": 0.17109180577099323,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.671509325504303,
+ "learning_rate": 3.5456828520060676e-06,
+ "loss": 0.10388398170471191,
+ "mean_token_accuracy": 0.9666652023792267,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.28785753225007754,
+ "eval_loss": 0.9345396161079407,
+ "eval_mean_token_accuracy": 0.8258346751164855,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 98.5634,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 3620
+ },
+ {
+ "entropy": 0.17080388870090246,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.9064943194389343,
+ "learning_rate": 3.2876680850949342e-06,
+ "loss": 0.09955486059188842,
+ "mean_token_accuracy": 0.9688052341341973,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.2863340745313784,
+ "eval_loss": 0.9368643760681152,
+ "eval_mean_token_accuracy": 0.826385034604019,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 98.4956,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 3640
+ },
+ {
+ "entropy": 0.1739195568487048,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.645410418510437,
+ "learning_rate": 3.0389516881466266e-06,
+ "loss": 0.1038020372390747,
+ "mean_token_accuracy": 0.965588228404522,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.2869644621282481,
+ "eval_loss": 0.9337670803070068,
+ "eval_mean_token_accuracy": 0.8266236021947325,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 98.3869,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 3660
+ },
+ {
+ "entropy": 0.16863423213362694,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.6110771298408508,
+ "learning_rate": 2.7996040449594445e-06,
+ "loss": 0.10157880783081055,
+ "mean_token_accuracy": 0.967547157406807,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.28859044577968257,
+ "eval_loss": 0.9297890067100525,
+ "eval_mean_token_accuracy": 0.8264903181724335,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 98.5054,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 3680
+ },
+ {
+ "entropy": 0.1767999932169914,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.6543433666229248,
+ "learning_rate": 2.5696928880852117e-06,
+ "loss": 0.10324364900588989,
+ "mean_token_accuracy": 0.9673267215490341,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.2885892426364877,
+ "eval_loss": 0.928764820098877,
+ "eval_mean_token_accuracy": 0.8265983676642514,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 98.4914,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 3700
+ },
+ {
+ "entropy": 0.16943628452718257,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.8403868079185486,
+ "learning_rate": 2.3492832796617217e-06,
+ "loss": 0.09948049783706665,
+ "mean_token_accuracy": 0.967706835269928,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.2884856275795551,
+ "eval_loss": 0.933276355266571,
+ "eval_mean_token_accuracy": 0.8261616119507993,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 98.6259,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 3720
+ },
+ {
+ "entropy": 0.17359357613783616,
+ "epoch": 9.012062726176115,
+ "grad_norm": 0.37213996052742004,
+ "learning_rate": 2.138437593000977e-06,
+ "loss": 0.10089871883392335,
+ "mean_token_accuracy": 0.9696638874518566,
+ "num_tokens": 9122597.0,
+ "step": 3740
+ },
+ {
+ "epoch": 9.012062726176115,
+ "eval_entropy": 0.2875421148505104,
+ "eval_loss": 0.9357590079307556,
+ "eval_mean_token_accuracy": 0.8264994859025719,
+ "eval_num_tokens": 9122597.0,
+ "eval_runtime": 98.523,
+ "eval_samples_per_second": 14.413,
+ "eval_steps_per_second": 1.807,
+ "step": 3740
+ },
+ {
+ "entropy": 0.1642639346420765,
+ "epoch": 9.060313630880579,
+ "grad_norm": 0.5511618852615356,
+ "learning_rate": 1.937215494938237e-06,
+ "loss": 0.08524587750434875,
+ "mean_token_accuracy": 0.9754035726189614,
+ "num_tokens": 9171477.0,
+ "step": 3760
+ },
+ {
+ "epoch": 9.060313630880579,
+ "eval_entropy": 0.2820046850469675,
+ "eval_loss": 0.9588018655776978,
+ "eval_mean_token_accuracy": 0.8257443647036392,
+ "eval_num_tokens": 9171477.0,
+ "eval_runtime": 98.3992,
+ "eval_samples_per_second": 14.431,
+ "eval_steps_per_second": 1.809,
+ "step": 3760
+ },
+ {
+ "entropy": 0.1630103051662445,
+ "epoch": 9.108564535585042,
+ "grad_norm": 0.5125291347503662,
+ "learning_rate": 1.7456739289470143e-06,
+ "loss": 0.0895907461643219,
+ "mean_token_accuracy": 0.973100857436657,
+ "num_tokens": 9217543.0,
+ "step": 3780
+ },
+ {
+ "epoch": 9.108564535585042,
+ "eval_entropy": 0.2786683516388529,
+ "eval_loss": 0.9743365049362183,
+ "eval_mean_token_accuracy": 0.8253507131940863,
+ "eval_num_tokens": 9217543.0,
+ "eval_runtime": 98.3882,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 3780
+ },
+ {
+ "entropy": 0.15873534101992845,
+ "epoch": 9.156815440289506,
+ "grad_norm": 0.5697775483131409,
+ "learning_rate": 1.5638670990247788e-06,
+ "loss": 0.08552997708320617,
+ "mean_token_accuracy": 0.973672679066658,
+ "num_tokens": 9268724.0,
+ "step": 3800
+ },
+ {
+ "epoch": 9.156815440289506,
+ "eval_entropy": 0.2781428360704626,
+ "eval_loss": 0.9790021777153015,
+ "eval_mean_token_accuracy": 0.8251898780967413,
+ "eval_num_tokens": 9268724.0,
+ "eval_runtime": 98.5449,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 3800
+ },
+ {
+ "entropy": 0.1606229657307267,
+ "epoch": 9.205066344993968,
+ "grad_norm": 0.5715742707252502,
+ "learning_rate": 1.3918464543538225e-06,
+ "loss": 0.08984670042991638,
+ "mean_token_accuracy": 0.9729854613542557,
+ "num_tokens": 9316805.0,
+ "step": 3820
+ },
+ {
+ "epoch": 9.205066344993968,
+ "eval_entropy": 0.2777416408899125,
+ "eval_loss": 0.9808777570724487,
+ "eval_mean_token_accuracy": 0.8252255926641185,
+ "eval_num_tokens": 9316805.0,
+ "eval_runtime": 98.627,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 3820
+ },
+ {
+ "entropy": 0.15027269292622805,
+ "epoch": 9.253317249698432,
+ "grad_norm": 0.3359828293323517,
+ "learning_rate": 1.229660674741786e-06,
+ "loss": 0.08229721784591675,
+ "mean_token_accuracy": 0.9754723474383354,
+ "num_tokens": 9369940.0,
+ "step": 3840
+ },
+ {
+ "epoch": 9.253317249698432,
+ "eval_entropy": 0.2768587637483404,
+ "eval_loss": 0.9837455749511719,
+ "eval_mean_token_accuracy": 0.825493957889214,
+ "eval_num_tokens": 9369940.0,
+ "eval_runtime": 98.534,
+ "eval_samples_per_second": 14.411,
+ "eval_steps_per_second": 1.806,
+ "step": 3840
+ },
+ {
+ "entropy": 0.16767099779099226,
+ "epoch": 9.301568154402895,
+ "grad_norm": 0.5708875060081482,
+ "learning_rate": 1.077355656845822e-06,
+ "loss": 0.08888025283813476,
+ "mean_token_accuracy": 0.9715678468346596,
+ "num_tokens": 9416788.0,
+ "step": 3860
+ },
+ {
+ "epoch": 9.301568154402895,
+ "eval_entropy": 0.27612990283229377,
+ "eval_loss": 0.9873999357223511,
+ "eval_mean_token_accuracy": 0.8252144948150335,
+ "eval_num_tokens": 9416788.0,
+ "eval_runtime": 98.3889,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 3860
+ },
+ {
+ "entropy": 0.1549418892711401,
+ "epoch": 9.349819059107359,
+ "grad_norm": 0.7601456046104431,
+ "learning_rate": 9.349745011844029e-07,
+ "loss": 0.08411069512367249,
+ "mean_token_accuracy": 0.9741103485226631,
+ "num_tokens": 9467862.0,
+ "step": 3880
+ },
+ {
+ "epoch": 9.349819059107359,
+ "eval_entropy": 0.2757877716689967,
+ "eval_loss": 0.9892058968544006,
+ "eval_mean_token_accuracy": 0.8250612017143978,
+ "eval_num_tokens": 9467862.0,
+ "eval_runtime": 98.3869,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 3880
+ },
+ {
+ "entropy": 0.16111517809331416,
+ "epoch": 9.398069963811821,
+ "grad_norm": 0.6774276494979858,
+ "learning_rate": 8.025574999403695e-07,
+ "loss": 0.08765592575073242,
+ "mean_token_accuracy": 0.9734174907207489,
+ "num_tokens": 9514611.0,
+ "step": 3900
+ },
+ {
+ "epoch": 9.398069963811821,
+ "eval_entropy": 0.275664389049739,
+ "eval_loss": 0.989257276058197,
+ "eval_mean_token_accuracy": 0.8252896914321385,
+ "eval_num_tokens": 9514611.0,
+ "eval_runtime": 98.4649,
+ "eval_samples_per_second": 14.421,
+ "eval_steps_per_second": 1.808,
+ "step": 3900
+ },
+ {
+ "entropy": 0.15701593235135078,
+ "epoch": 9.446320868516285,
+ "grad_norm": 0.6130298376083374,
+ "learning_rate": 6.801421255587377e-07,
+ "loss": 0.0853091835975647,
+ "mean_token_accuracy": 0.9740088582038879,
+ "num_tokens": 9563437.0,
+ "step": 3920
+ },
+ {
+ "epoch": 9.446320868516285,
+ "eval_entropy": 0.27559053881114787,
+ "eval_loss": 0.9901270866394043,
+ "eval_mean_token_accuracy": 0.8251507653948966,
+ "eval_num_tokens": 9563437.0,
+ "eval_runtime": 98.5285,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 3920
+ },
+ {
+ "entropy": 0.15848954878747462,
+ "epoch": 9.494571773220748,
+ "grad_norm": 0.5928868651390076,
+ "learning_rate": 5.677630201424204e-07,
+ "loss": 0.08467000126838684,
+ "mean_token_accuracy": 0.9725634664297104,
+ "num_tokens": 9614034.0,
+ "step": 3940
+ },
+ {
+ "epoch": 9.494571773220748,
+ "eval_entropy": 0.2750965692521481,
+ "eval_loss": 0.9914345145225525,
+ "eval_mean_token_accuracy": 0.8252089954494091,
+ "eval_num_tokens": 9614034.0,
+ "eval_runtime": 98.5049,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 3940
+ },
+ {
+ "entropy": 0.15644967313855887,
+ "epoch": 9.54282267792521,
+ "grad_norm": 0.5718610882759094,
+ "learning_rate": 4.6545198564893463e-07,
+ "loss": 0.08564762473106384,
+ "mean_token_accuracy": 0.9730244860053062,
+ "num_tokens": 9661266.0,
+ "step": 3960
+ },
+ {
+ "epoch": 9.54282267792521,
+ "eval_entropy": 0.27522430820076654,
+ "eval_loss": 0.9908939003944397,
+ "eval_mean_token_accuracy": 0.8251052994406625,
+ "eval_num_tokens": 9661266.0,
+ "eval_runtime": 98.3637,
+ "eval_samples_per_second": 14.436,
+ "eval_steps_per_second": 1.81,
+ "step": 3960
+ },
+ {
+ "entropy": 0.15196608882397414,
+ "epoch": 9.591073582629674,
+ "grad_norm": 0.752661406993866,
+ "learning_rate": 3.732379748908161e-07,
+ "loss": 0.08072127103805542,
+ "mean_token_accuracy": 0.9749975100159645,
+ "num_tokens": 9710458.0,
+ "step": 3980
+ },
+ {
+ "epoch": 9.591073582629674,
+ "eval_entropy": 0.2750371985890892,
+ "eval_loss": 0.9913897514343262,
+ "eval_mean_token_accuracy": 0.8248539731073915,
+ "eval_num_tokens": 9710458.0,
+ "eval_runtime": 98.4959,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 3980
+ },
+ {
+ "entropy": 0.16616767533123494,
+ "epoch": 9.639324487334138,
+ "grad_norm": 0.7107775211334229,
+ "learning_rate": 2.911470833423274e-07,
+ "loss": 0.08988997340202332,
+ "mean_token_accuracy": 0.9714749112725258,
+ "num_tokens": 9755797.0,
+ "step": 4000
+ },
+ {
+ "epoch": 9.639324487334138,
+ "eval_entropy": 0.27523158457172053,
+ "eval_loss": 0.9913759827613831,
+ "eval_mean_token_accuracy": 0.8250464505693885,
+ "eval_num_tokens": 9755797.0,
+ "eval_runtime": 98.5067,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 4000
+ },
+ {
+ "entropy": 0.15760798901319503,
+ "epoch": 9.687575392038601,
+ "grad_norm": 0.5563387870788574,
+ "learning_rate": 2.1920254175472834e-07,
+ "loss": 0.0877057135105133,
+ "mean_token_accuracy": 0.9734948605298996,
+ "num_tokens": 9805552.0,
+ "step": 4020
+ },
+ {
+ "epoch": 9.687575392038601,
+ "eval_entropy": 0.27528667726208655,
+ "eval_loss": 0.9914706349372864,
+ "eval_mean_token_accuracy": 0.8250808802883277,
+ "eval_num_tokens": 9805552.0,
+ "eval_runtime": 98.475,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 4020
+ },
+ {
+ "entropy": 0.161653464846313,
+ "epoch": 9.735826296743063,
+ "grad_norm": 0.5896415114402771,
+ "learning_rate": 1.5742470958227596e-07,
+ "loss": 0.08775997757911683,
+ "mean_token_accuracy": 0.9718310445547104,
+ "num_tokens": 9852843.0,
+ "step": 4040
+ },
+ {
+ "epoch": 9.735826296743063,
+ "eval_entropy": 0.27531274182073184,
+ "eval_loss": 0.9915319085121155,
+ "eval_mean_token_accuracy": 0.825061980258213,
+ "eval_num_tokens": 9852843.0,
+ "eval_runtime": 98.3873,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 4040
+ },
+ {
+ "entropy": 0.16010624952614308,
+ "epoch": 9.784077201447527,
+ "grad_norm": 0.6022347211837769,
+ "learning_rate": 1.058310692207063e-07,
+ "loss": 0.08761411905288696,
+ "mean_token_accuracy": 0.9729624018073082,
+ "num_tokens": 9900151.0,
+ "step": 4060
+ },
+ {
+ "epoch": 9.784077201447527,
+ "eval_entropy": 0.27524732446737504,
+ "eval_loss": 0.991264820098877,
+ "eval_mean_token_accuracy": 0.8251047007153544,
+ "eval_num_tokens": 9900151.0,
+ "eval_runtime": 98.4834,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 4060
+ },
+ {
+ "entropy": 0.159021489135921,
+ "epoch": 9.83232810615199,
+ "grad_norm": 0.48451653122901917,
+ "learning_rate": 6.443622105994901e-08,
+ "loss": 0.08291574716567993,
+ "mean_token_accuracy": 0.9726693674921989,
+ "num_tokens": 9952926.0,
+ "step": 4080
+ },
+ {
+ "epoch": 9.83232810615199,
+ "eval_entropy": 0.27521907144717955,
+ "eval_loss": 0.9918096661567688,
+ "eval_mean_token_accuracy": 0.8250070894032382,
+ "eval_num_tokens": 9952926.0,
+ "eval_runtime": 98.3568,
+ "eval_samples_per_second": 14.437,
+ "eval_steps_per_second": 1.81,
+ "step": 4080
+ },
+ {
+ "entropy": 0.15175675600767136,
+ "epoch": 9.880579010856454,
+ "grad_norm": 0.5009680390357971,
+ "learning_rate": 3.325187935238231e-08,
+ "loss": 0.08214685916900635,
+ "mean_token_accuracy": 0.9752551406621933,
+ "num_tokens": 10003720.0,
+ "step": 4100
+ },
+ {
+ "epoch": 9.880579010856454,
+ "eval_entropy": 0.27522189488236826,
+ "eval_loss": 0.9918575882911682,
+ "eval_mean_token_accuracy": 0.8249149091458052,
+ "eval_num_tokens": 10003720.0,
+ "eval_runtime": 98.4802,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 4100
+ },
+ {
+ "entropy": 0.15018076542764902,
+ "epoch": 9.928829915560916,
+ "grad_norm": 0.6191889047622681,
+ "learning_rate": 1.2286868897828347e-08,
+ "loss": 0.08095031380653381,
+ "mean_token_accuracy": 0.9739995509386062,
+ "num_tokens": 10057590.0,
+ "step": 4120
+ },
+ {
+ "epoch": 9.928829915560916,
+ "eval_entropy": 0.27517637836464337,
+ "eval_loss": 0.991787850856781,
+ "eval_mean_token_accuracy": 0.8249709994605418,
+ "eval_num_tokens": 10057590.0,
+ "eval_runtime": 98.5572,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 4120
+ },
+ {
+ "entropy": 0.16240970119833947,
+ "epoch": 9.97708082026538,
+ "grad_norm": 0.5195524096488953,
+ "learning_rate": 1.5471225462640753e-09,
+ "loss": 0.08831822276115417,
+ "mean_token_accuracy": 0.9716430872678756,
+ "num_tokens": 10103443.0,
+ "step": 4140
+ },
+ {
+ "epoch": 9.97708082026538,
+ "eval_entropy": 0.27521504202250685,
+ "eval_loss": 0.9918516278266907,
+ "eval_mean_token_accuracy": 0.8251796190658313,
+ "eval_num_tokens": 10103443.0,
+ "eval_runtime": 98.3776,
+ "eval_samples_per_second": 14.434,
+ "eval_steps_per_second": 1.809,
+ "step": 4140
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 2.094999795151872e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4150/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4150/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4150/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4150/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4150/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4150/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4150/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4150/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4150/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4150/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4150/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4150/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4150/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4150/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..e7e2051912a2f1dbcf2d87f1f301dfe65bba6f09
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-4150/trainer_state.json
@@ -0,0 +1,4392 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 10.0,
+ "eval_steps": 20,
+ "global_step": 4150,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ },
+ {
+ "entropy": 0.6141144648194313,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6467089056968689,
+ "learning_rate": 7.078840074092574e-05,
+ "loss": 0.5548617362976074,
+ "mean_token_accuracy": 0.8342376738786698,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.5951976082967908,
+ "eval_loss": 0.5850171446800232,
+ "eval_mean_token_accuracy": 0.8303041870004675,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 98.6601,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 760
+ },
+ {
+ "entropy": 0.5961867064237595,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.41280242800712585,
+ "learning_rate": 7.061001336495465e-05,
+ "loss": 0.5511267185211182,
+ "mean_token_accuracy": 0.8392465204000473,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6167017998320333,
+ "eval_loss": 0.5826791524887085,
+ "eval_mean_token_accuracy": 0.8309678050239434,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 98.6564,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 780
+ },
+ {
+ "entropy": 0.6025628983974457,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.4078947901725769,
+ "learning_rate": 7.042187293157881e-05,
+ "loss": 0.5552026271820069,
+ "mean_token_accuracy": 0.8359778299927711,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.6047008484601974,
+ "eval_loss": 0.5811123847961426,
+ "eval_mean_token_accuracy": 0.8320730804057603,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 98.6033,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 800
+ },
+ {
+ "entropy": 0.6041995719075203,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.543890118598938,
+ "learning_rate": 7.022403268231535e-05,
+ "loss": 0.554384422302246,
+ "mean_token_accuracy": 0.8380729153752327,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.6005127917179901,
+ "eval_loss": 0.5800131559371948,
+ "eval_mean_token_accuracy": 0.8313186547059691,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 98.4119,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 820
+ },
+ {
+ "entropy": 0.567442889396961,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.5015465021133423,
+ "learning_rate": 7.001654860361456e-05,
+ "loss": 0.5208251476287842,
+ "mean_token_accuracy": 0.8442784960453327,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.5785037809878253,
+ "eval_loss": 0.5865591168403625,
+ "eval_mean_token_accuracy": 0.8307993857378371,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 98.5061,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 840
+ },
+ {
+ "entropy": 0.5552012033760547,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.5975943207740784,
+ "learning_rate": 6.979947941101646e-05,
+ "loss": 0.5064489841461182,
+ "mean_token_accuracy": 0.8468396887183189,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5844205239515626,
+ "eval_loss": 0.585284948348999,
+ "eval_mean_token_accuracy": 0.8309797065311604,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 98.4982,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 860
+ },
+ {
+ "entropy": 0.5401638634502888,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.4194582998752594,
+ "learning_rate": 6.957288653253486e-05,
+ "loss": 0.49037814140319824,
+ "mean_token_accuracy": 0.8528496414422989,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.574507936835289,
+ "eval_loss": 0.5836747288703918,
+ "eval_mean_token_accuracy": 0.8325241643391298,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 98.5592,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 880
+ },
+ {
+ "entropy": 0.5606910541653634,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5395355820655823,
+ "learning_rate": 6.933683409127401e-05,
+ "loss": 0.5089571475982666,
+ "mean_token_accuracy": 0.8459660410881042,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5779740023479033,
+ "eval_loss": 0.5816043615341187,
+ "eval_mean_token_accuracy": 0.8326318036304431,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 98.4386,
+ "eval_samples_per_second": 14.425,
+ "eval_steps_per_second": 1.808,
+ "step": 900
+ },
+ {
+ "entropy": 0.5380903735756875,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.46588417887687683,
+ "learning_rate": 6.909138888728257e-05,
+ "loss": 0.4902662754058838,
+ "mean_token_accuracy": 0.8493925213813782,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.570154261220707,
+ "eval_loss": 0.5827520489692688,
+ "eval_mean_token_accuracy": 0.8321914167216654,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 98.5674,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 920
+ },
+ {
+ "entropy": 0.558555257320404,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.44576048851013184,
+ "learning_rate": 6.883662037864996e-05,
+ "loss": 0.5000699520111084,
+ "mean_token_accuracy": 0.8475695163011551,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5715417749761196,
+ "eval_loss": 0.5831062197685242,
+ "eval_mean_token_accuracy": 0.8327409899636601,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 98.568,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 940
+ },
+ {
+ "entropy": 0.5582522705197335,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.38966405391693115,
+ "learning_rate": 6.857260066185045e-05,
+ "loss": 0.5132341861724854,
+ "mean_token_accuracy": 0.8466864034533501,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5823918579334624,
+ "eval_loss": 0.5797588229179382,
+ "eval_mean_token_accuracy": 0.8326037566313583,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 98.6222,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 960
+ },
+ {
+ "entropy": 0.5557353265583516,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.42300811409950256,
+ "learning_rate": 6.829940445134089e-05,
+ "loss": 0.5035375118255615,
+ "mean_token_accuracy": 0.8479988545179367,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5542296482605881,
+ "eval_loss": 0.5800604820251465,
+ "eval_mean_token_accuracy": 0.8336404163515969,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 98.615,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 980
+ },
+ {
+ "entropy": 0.5492161035537719,
+ "epoch": 2.4101326899879374,
+ "grad_norm": 0.47482752799987793,
+ "learning_rate": 6.801710905841724e-05,
+ "loss": 0.5032052040100098,
+ "mean_token_accuracy": 0.8488596931099892,
+ "num_tokens": 2443253.0,
+ "step": 1000
+ },
+ {
+ "epoch": 2.4101326899879374,
+ "eval_entropy": 0.5755258423558781,
+ "eval_loss": 0.5764774680137634,
+ "eval_mean_token_accuracy": 0.8336571763070781,
+ "eval_num_tokens": 2443253.0,
+ "eval_runtime": 98.4765,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 1000
+ },
+ {
+ "entropy": 0.5439704649150372,
+ "epoch": 2.4583835946924006,
+ "grad_norm": 0.3656543493270874,
+ "learning_rate": 6.772579436933644e-05,
+ "loss": 0.49903340339660646,
+ "mean_token_accuracy": 0.8499931409955025,
+ "num_tokens": 2495178.0,
+ "step": 1020
+ },
+ {
+ "epoch": 2.4583835946924006,
+ "eval_entropy": 0.5712662409530597,
+ "eval_loss": 0.5777108073234558,
+ "eval_mean_token_accuracy": 0.8336941352051296,
+ "eval_num_tokens": 2495178.0,
+ "eval_runtime": 98.678,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 1020
+ },
+ {
+ "entropy": 0.5448883637785912,
+ "epoch": 2.5066344993968634,
+ "grad_norm": 0.3990268111228943,
+ "learning_rate": 6.742554282270959e-05,
+ "loss": 0.4990490436553955,
+ "mean_token_accuracy": 0.8488436549901962,
+ "num_tokens": 2541733.0,
+ "step": 1040
+ },
+ {
+ "epoch": 2.5066344993968634,
+ "eval_entropy": 0.5816385483139017,
+ "eval_loss": 0.5749375820159912,
+ "eval_mean_token_accuracy": 0.833801462074344,
+ "eval_num_tokens": 2541733.0,
+ "eval_runtime": 98.6209,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 1040
+ },
+ {
+ "entropy": 0.5575303137302399,
+ "epoch": 2.554885404101327,
+ "grad_norm": 0.45015832781791687,
+ "learning_rate": 6.711643938617279e-05,
+ "loss": 0.5061063289642334,
+ "mean_token_accuracy": 0.8462719902396202,
+ "num_tokens": 2589479.0,
+ "step": 1060
+ },
+ {
+ "epoch": 2.554885404101327,
+ "eval_entropy": 0.5666238901655326,
+ "eval_loss": 0.574820876121521,
+ "eval_mean_token_accuracy": 0.8339746617869045,
+ "eval_num_tokens": 2589479.0,
+ "eval_runtime": 98.5332,
+ "eval_samples_per_second": 14.411,
+ "eval_steps_per_second": 1.806,
+ "step": 1060
+ },
+ {
+ "entropy": 0.5477629877626896,
+ "epoch": 2.60313630880579,
+ "grad_norm": 0.43583670258522034,
+ "learning_rate": 6.679857153234218e-05,
+ "loss": 0.4998013496398926,
+ "mean_token_accuracy": 0.8465577483177185,
+ "num_tokens": 2639134.0,
+ "step": 1080
+ },
+ {
+ "epoch": 2.60313630880579,
+ "eval_entropy": 0.5809191281206152,
+ "eval_loss": 0.5737497210502625,
+ "eval_mean_token_accuracy": 0.8338909467284599,
+ "eval_num_tokens": 2639134.0,
+ "eval_runtime": 98.5093,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 1080
+ },
+ {
+ "entropy": 0.5548589795827865,
+ "epoch": 2.651387213510253,
+ "grad_norm": 0.44145169854164124,
+ "learning_rate": 6.647202921406043e-05,
+ "loss": 0.5035099029541016,
+ "mean_token_accuracy": 0.8463503196835518,
+ "num_tokens": 2689324.0,
+ "step": 1100
+ },
+ {
+ "epoch": 2.651387213510253,
+ "eval_entropy": 0.5626686118292005,
+ "eval_loss": 0.5707846879959106,
+ "eval_mean_token_accuracy": 0.8349729531936432,
+ "eval_num_tokens": 2689324.0,
+ "eval_runtime": 98.473,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 1100
+ },
+ {
+ "entropy": 0.5568635657429695,
+ "epoch": 2.6996381182147164,
+ "grad_norm": 0.38482189178466797,
+ "learning_rate": 6.613690483894096e-05,
+ "loss": 0.5090720176696777,
+ "mean_token_accuracy": 0.8481037884950637,
+ "num_tokens": 2736182.0,
+ "step": 1120
+ },
+ {
+ "epoch": 2.6996381182147164,
+ "eval_entropy": 0.5752792709998871,
+ "eval_loss": 0.5706034898757935,
+ "eval_mean_token_accuracy": 0.8350465498613507,
+ "eval_num_tokens": 2736182.0,
+ "eval_runtime": 98.6524,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 1120
+ },
+ {
+ "entropy": 0.5644363485276699,
+ "epoch": 2.7478890229191797,
+ "grad_norm": 0.4353690445423126,
+ "learning_rate": 6.57932932432178e-05,
+ "loss": 0.5078121185302734,
+ "mean_token_accuracy": 0.846393559873104,
+ "num_tokens": 2781503.0,
+ "step": 1140
+ },
+ {
+ "epoch": 2.7478890229191797,
+ "eval_entropy": 0.5670730096570561,
+ "eval_loss": 0.5699080228805542,
+ "eval_mean_token_accuracy": 0.8354134288396728,
+ "eval_num_tokens": 2781503.0,
+ "eval_runtime": 98.5697,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 1140
+ },
+ {
+ "entropy": 0.547992604970932,
+ "epoch": 2.796139927623643,
+ "grad_norm": 0.4126238524913788,
+ "learning_rate": 6.544129166490802e-05,
+ "loss": 0.4943515777587891,
+ "mean_token_accuracy": 0.8496095523238182,
+ "num_tokens": 2831126.0,
+ "step": 1160
+ },
+ {
+ "epoch": 2.796139927623643,
+ "eval_entropy": 0.5552005295673114,
+ "eval_loss": 0.5671029090881348,
+ "eval_mean_token_accuracy": 0.8364275894807965,
+ "eval_num_tokens": 2831126.0,
+ "eval_runtime": 98.5412,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 1160
+ },
+ {
+ "entropy": 0.561221782118082,
+ "epoch": 2.844390832328106,
+ "grad_norm": 0.47753170132637024,
+ "learning_rate": 6.508099971629446e-05,
+ "loss": 0.5094280242919922,
+ "mean_token_accuracy": 0.8462468951940536,
+ "num_tokens": 2878560.0,
+ "step": 1180
+ },
+ {
+ "epoch": 2.844390832328106,
+ "eval_entropy": 0.5532578974962234,
+ "eval_loss": 0.5714129209518433,
+ "eval_mean_token_accuracy": 0.8357450737712089,
+ "eval_num_tokens": 2878560.0,
+ "eval_runtime": 98.5358,
+ "eval_samples_per_second": 14.411,
+ "eval_steps_per_second": 1.806,
+ "step": 1180
+ },
+ {
+ "entropy": 0.5400963060557842,
+ "epoch": 2.8926417370325694,
+ "grad_norm": 0.38486021757125854,
+ "learning_rate": 6.471251935573679e-05,
+ "loss": 0.49443836212158204,
+ "mean_token_accuracy": 0.8484652638435364,
+ "num_tokens": 2927569.0,
+ "step": 1200
+ },
+ {
+ "epoch": 2.8926417370325694,
+ "eval_entropy": 0.5697383115465722,
+ "eval_loss": 0.5681865215301514,
+ "eval_mean_token_accuracy": 0.8350923453154189,
+ "eval_num_tokens": 2927569.0,
+ "eval_runtime": 98.4136,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 1200
+ },
+ {
+ "entropy": 0.5381059750914574,
+ "epoch": 2.9408926417370327,
+ "grad_norm": 0.49116969108581543,
+ "learning_rate": 6.43359548588183e-05,
+ "loss": 0.4878206729888916,
+ "mean_token_accuracy": 0.8522022038698196,
+ "num_tokens": 2978577.0,
+ "step": 1220
+ },
+ {
+ "epoch": 2.9408926417370327,
+ "eval_entropy": 0.5614734472853414,
+ "eval_loss": 0.5665943622589111,
+ "eval_mean_token_accuracy": 0.8360684296388304,
+ "eval_num_tokens": 2978577.0,
+ "eval_runtime": 98.5682,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 1220
+ },
+ {
+ "entropy": 0.5569418989121914,
+ "epoch": 2.989143546441496,
+ "grad_norm": 0.39628803730010986,
+ "learning_rate": 6.39514127888373e-05,
+ "loss": 0.5087827682495117,
+ "mean_token_accuracy": 0.8476181954145432,
+ "num_tokens": 3025565.0,
+ "step": 1240
+ },
+ {
+ "epoch": 2.989143546441496,
+ "eval_entropy": 0.5740904657358534,
+ "eval_loss": 0.5637108683586121,
+ "eval_mean_token_accuracy": 0.8366153638684348,
+ "eval_num_tokens": 3025565.0,
+ "eval_runtime": 98.4904,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 1240
+ },
+ {
+ "entropy": 0.5067426371268737,
+ "epoch": 3.0361881785283473,
+ "grad_norm": 0.4593591094017029,
+ "learning_rate": 6.355900196665094e-05,
+ "loss": 0.4515994071960449,
+ "mean_token_accuracy": 0.8621168640943674,
+ "num_tokens": 3073705.0,
+ "step": 1260
+ },
+ {
+ "epoch": 3.0361881785283473,
+ "eval_entropy": 0.5176989454566763,
+ "eval_loss": 0.5825926661491394,
+ "eval_mean_token_accuracy": 0.8353930775369152,
+ "eval_num_tokens": 3073705.0,
+ "eval_runtime": 98.5654,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 1260
+ },
+ {
+ "entropy": 0.4712781891226768,
+ "epoch": 3.0844390832328106,
+ "grad_norm": 0.46717822551727295,
+ "learning_rate": 6.315883343988008e-05,
+ "loss": 0.4228676795959473,
+ "mean_token_accuracy": 0.8672123551368713,
+ "num_tokens": 3126555.0,
+ "step": 1280
+ },
+ {
+ "epoch": 3.0844390832328106,
+ "eval_entropy": 0.5229383730486538,
+ "eval_loss": 0.5875843167304993,
+ "eval_mean_token_accuracy": 0.8347370524084969,
+ "eval_num_tokens": 3126555.0,
+ "eval_runtime": 98.5459,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 1280
+ },
+ {
+ "entropy": 0.48235836550593375,
+ "epoch": 3.132689987937274,
+ "grad_norm": 0.5073364973068237,
+ "learning_rate": 6.275102045148425e-05,
+ "loss": 0.4365375518798828,
+ "mean_token_accuracy": 0.8644895792007447,
+ "num_tokens": 3179060.0,
+ "step": 1300
+ },
+ {
+ "epoch": 3.132689987937274,
+ "eval_entropy": 0.5151654947339819,
+ "eval_loss": 0.5826129913330078,
+ "eval_mean_token_accuracy": 0.8359152366606037,
+ "eval_num_tokens": 3179060.0,
+ "eval_runtime": 98.3713,
+ "eval_samples_per_second": 14.435,
+ "eval_steps_per_second": 1.809,
+ "step": 1300
+ },
+ {
+ "entropy": 0.4756111137568951,
+ "epoch": 3.180940892641737,
+ "grad_norm": 0.4896017909049988,
+ "learning_rate": 6.233567840771526e-05,
+ "loss": 0.43177642822265627,
+ "mean_token_accuracy": 0.8645799145102501,
+ "num_tokens": 3234212.0,
+ "step": 1320
+ },
+ {
+ "epoch": 3.180940892641737,
+ "eval_entropy": 0.5227621813838401,
+ "eval_loss": 0.5796895027160645,
+ "eval_mean_token_accuracy": 0.8355028448479899,
+ "eval_num_tokens": 3234212.0,
+ "eval_runtime": 98.5538,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 1320
+ },
+ {
+ "entropy": 0.47541345208883284,
+ "epoch": 3.2291917973462003,
+ "grad_norm": 0.5064273476600647,
+ "learning_rate": 6.191292484545842e-05,
+ "loss": 0.4289750099182129,
+ "mean_token_accuracy": 0.8659383848309516,
+ "num_tokens": 3282544.0,
+ "step": 1340
+ },
+ {
+ "epoch": 3.2291917973462003,
+ "eval_entropy": 0.5286844237132019,
+ "eval_loss": 0.579947829246521,
+ "eval_mean_token_accuracy": 0.8353635648663125,
+ "eval_num_tokens": 3282544.0,
+ "eval_runtime": 98.589,
+ "eval_samples_per_second": 14.403,
+ "eval_steps_per_second": 1.805,
+ "step": 1340
+ },
+ {
+ "entropy": 0.4916382558643818,
+ "epoch": 3.2774427020506636,
+ "grad_norm": 0.5225116014480591,
+ "learning_rate": 6.148287939897118e-05,
+ "loss": 0.44352192878723146,
+ "mean_token_accuracy": 0.8630174115300179,
+ "num_tokens": 3332532.0,
+ "step": 1360
+ },
+ {
+ "epoch": 3.2774427020506636,
+ "eval_entropy": 0.5264859509267165,
+ "eval_loss": 0.5774893164634705,
+ "eval_mean_token_accuracy": 0.8363003827882617,
+ "eval_num_tokens": 3332532.0,
+ "eval_runtime": 98.5858,
+ "eval_samples_per_second": 14.404,
+ "eval_steps_per_second": 1.806,
+ "step": 1360
+ },
+ {
+ "entropy": 0.4990046709775925,
+ "epoch": 3.325693606755127,
+ "grad_norm": 0.548435628414154,
+ "learning_rate": 6.104566376602795e-05,
+ "loss": 0.4463675498962402,
+ "mean_token_accuracy": 0.8630342438817025,
+ "num_tokens": 3378178.0,
+ "step": 1380
+ },
+ {
+ "epoch": 3.325693606755127,
+ "eval_entropy": 0.522659856616781,
+ "eval_loss": 0.5790286660194397,
+ "eval_mean_token_accuracy": 0.8355394277010071,
+ "eval_num_tokens": 3378178.0,
+ "eval_runtime": 98.4707,
+ "eval_samples_per_second": 14.421,
+ "eval_steps_per_second": 1.808,
+ "step": 1380
+ },
+ {
+ "entropy": 0.48943497613072395,
+ "epoch": 3.37394451145959,
+ "grad_norm": 0.6222821474075317,
+ "learning_rate": 6.060140167348109e-05,
+ "loss": 0.4356812000274658,
+ "mean_token_accuracy": 0.8644352272152901,
+ "num_tokens": 3425739.0,
+ "step": 1400
+ },
+ {
+ "epoch": 3.37394451145959,
+ "eval_entropy": 0.5065617703654793,
+ "eval_loss": 0.5842992067337036,
+ "eval_mean_token_accuracy": 0.8358401662847968,
+ "eval_num_tokens": 3425739.0,
+ "eval_runtime": 98.4775,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 1400
+ },
+ {
+ "entropy": 0.47648413479328156,
+ "epoch": 3.422195416164053,
+ "grad_norm": 0.5207330584526062,
+ "learning_rate": 6.015021884224766e-05,
+ "loss": 0.42724952697753904,
+ "mean_token_accuracy": 0.8663465172052384,
+ "num_tokens": 3472737.0,
+ "step": 1420
+ },
+ {
+ "epoch": 3.422195416164053,
+ "eval_entropy": 0.5232780023237292,
+ "eval_loss": 0.5769163966178894,
+ "eval_mean_token_accuracy": 0.8356800042511372,
+ "eval_num_tokens": 3472737.0,
+ "eval_runtime": 98.5005,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 1420
+ },
+ {
+ "entropy": 0.4767202906310558,
+ "epoch": 3.470446320868516,
+ "grad_norm": 0.4695364534854889,
+ "learning_rate": 5.969224295173193e-05,
+ "loss": 0.42672252655029297,
+ "mean_token_accuracy": 0.8657647699117661,
+ "num_tokens": 3519802.0,
+ "step": 1440
+ },
+ {
+ "epoch": 3.470446320868516,
+ "eval_entropy": 0.5146279760291067,
+ "eval_loss": 0.577904224395752,
+ "eval_mean_token_accuracy": 0.8365512393163831,
+ "eval_num_tokens": 3519802.0,
+ "eval_runtime": 98.5825,
+ "eval_samples_per_second": 14.404,
+ "eval_steps_per_second": 1.806,
+ "step": 1440
+ },
+ {
+ "entropy": 0.4855267658829689,
+ "epoch": 3.5186972255729794,
+ "grad_norm": 0.6333373785018921,
+ "learning_rate": 5.92276036036936e-05,
+ "loss": 0.4310962677001953,
+ "mean_token_accuracy": 0.864738130569458,
+ "num_tokens": 3568814.0,
+ "step": 1460
+ },
+ {
+ "epoch": 3.5186972255729794,
+ "eval_entropy": 0.5260473481389913,
+ "eval_loss": 0.5770025253295898,
+ "eval_mean_token_accuracy": 0.8357346339172191,
+ "eval_num_tokens": 3568814.0,
+ "eval_runtime": 98.462,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 1460
+ },
+ {
+ "entropy": 0.4917050123214722,
+ "epoch": 3.5669481302774426,
+ "grad_norm": 0.5833342671394348,
+ "learning_rate": 5.875643228557199e-05,
+ "loss": 0.44533047676086424,
+ "mean_token_accuracy": 0.8642907112836837,
+ "num_tokens": 3616502.0,
+ "step": 1480
+ },
+ {
+ "epoch": 3.5669481302774426,
+ "eval_entropy": 0.5278292167722509,
+ "eval_loss": 0.5773910880088806,
+ "eval_mean_token_accuracy": 0.83580608548743,
+ "eval_num_tokens": 3616502.0,
+ "eval_runtime": 98.4817,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 1480
+ },
+ {
+ "entropy": 0.5025837384164333,
+ "epoch": 3.615199034981906,
+ "grad_norm": 0.5484253168106079,
+ "learning_rate": 5.827886233327671e-05,
+ "loss": 0.44471564292907717,
+ "mean_token_accuracy": 0.8624475196003913,
+ "num_tokens": 3663750.0,
+ "step": 1500
+ },
+ {
+ "epoch": 3.615199034981906,
+ "eval_entropy": 0.5118448720889145,
+ "eval_loss": 0.5786464810371399,
+ "eval_mean_token_accuracy": 0.8370849305994055,
+ "eval_num_tokens": 3663750.0,
+ "eval_runtime": 98.5208,
+ "eval_samples_per_second": 14.413,
+ "eval_steps_per_second": 1.807,
+ "step": 1500
+ },
+ {
+ "entropy": 0.494297581911087,
+ "epoch": 3.663449939686369,
+ "grad_norm": 0.5475512146949768,
+ "learning_rate": 5.7795028893455165e-05,
+ "loss": 0.44118366241455076,
+ "mean_token_accuracy": 0.8630154177546501,
+ "num_tokens": 3709337.0,
+ "step": 1520
+ },
+ {
+ "epoch": 3.663449939686369,
+ "eval_entropy": 0.5250470112548785,
+ "eval_loss": 0.5746602416038513,
+ "eval_mean_token_accuracy": 0.8370710711800651,
+ "eval_num_tokens": 3709337.0,
+ "eval_runtime": 98.54,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 1520
+ },
+ {
+ "entropy": 0.4840923048555851,
+ "epoch": 3.7117008443908324,
+ "grad_norm": 0.39450952410697937,
+ "learning_rate": 5.730506888524761e-05,
+ "loss": 0.443526554107666,
+ "mean_token_accuracy": 0.8627525731921196,
+ "num_tokens": 3759111.0,
+ "step": 1540
+ },
+ {
+ "epoch": 3.7117008443908324,
+ "eval_entropy": 0.5305671740448876,
+ "eval_loss": 0.5697008371353149,
+ "eval_mean_token_accuracy": 0.837059904015466,
+ "eval_num_tokens": 3759111.0,
+ "eval_runtime": 98.6288,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 1540
+ },
+ {
+ "entropy": 0.49613324701786043,
+ "epoch": 3.7599517490952956,
+ "grad_norm": 0.47117286920547485,
+ "learning_rate": 5.680912096154071e-05,
+ "loss": 0.44313778877258303,
+ "mean_token_accuracy": 0.8614323407411575,
+ "num_tokens": 3806869.0,
+ "step": 1560
+ },
+ {
+ "epoch": 3.7599517490952956,
+ "eval_entropy": 0.522762700580479,
+ "eval_loss": 0.571703314781189,
+ "eval_mean_token_accuracy": 0.8381530970669864,
+ "eval_num_tokens": 3806869.0,
+ "eval_runtime": 98.6493,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 1560
+ },
+ {
+ "entropy": 0.4979200787842274,
+ "epoch": 3.808202653799759,
+ "grad_norm": 0.5275628566741943,
+ "learning_rate": 5.630732546973038e-05,
+ "loss": 0.4462272167205811,
+ "mean_token_accuracy": 0.8615201056003571,
+ "num_tokens": 3854699.0,
+ "step": 1580
+ },
+ {
+ "epoch": 3.808202653799759,
+ "eval_entropy": 0.5232290047942922,
+ "eval_loss": 0.568954586982727,
+ "eval_mean_token_accuracy": 0.8380421128835571,
+ "eval_num_tokens": 3854699.0,
+ "eval_runtime": 98.484,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 1580
+ },
+ {
+ "entropy": 0.4957688055932522,
+ "epoch": 3.856453558504222,
+ "grad_norm": 0.5806206464767456,
+ "learning_rate": 5.5799824412005126e-05,
+ "loss": 0.44518418312072755,
+ "mean_token_accuracy": 0.860895662009716,
+ "num_tokens": 3903750.0,
+ "step": 1600
+ },
+ {
+ "epoch": 3.856453558504222,
+ "eval_entropy": 0.5238539929805177,
+ "eval_loss": 0.573081910610199,
+ "eval_mean_token_accuracy": 0.8369396288743179,
+ "eval_num_tokens": 3903750.0,
+ "eval_runtime": 98.5731,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 1600
+ },
+ {
+ "entropy": 0.5005358196794987,
+ "epoch": 3.9047044632086854,
+ "grad_norm": 0.5114495754241943,
+ "learning_rate": 5.528676140516111e-05,
+ "loss": 0.45078649520874026,
+ "mean_token_accuracy": 0.860628567636013,
+ "num_tokens": 3952913.0,
+ "step": 1620
+ },
+ {
+ "epoch": 3.9047044632086854,
+ "eval_entropy": 0.5245393164707034,
+ "eval_loss": 0.5683825612068176,
+ "eval_mean_token_accuracy": 0.8380641599049728,
+ "eval_num_tokens": 3952913.0,
+ "eval_runtime": 98.5275,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 1620
+ },
+ {
+ "entropy": 0.4839292488992214,
+ "epoch": 3.952955367913148,
+ "grad_norm": 0.6169933080673218,
+ "learning_rate": 5.4768281639960344e-05,
+ "loss": 0.43407382965087893,
+ "mean_token_accuracy": 0.867341585457325,
+ "num_tokens": 4002571.0,
+ "step": 1640
+ },
+ {
+ "epoch": 3.952955367913148,
+ "eval_entropy": 0.5277225765284528,
+ "eval_loss": 0.567344069480896,
+ "eval_mean_token_accuracy": 0.8386145431673928,
+ "eval_num_tokens": 4002571.0,
+ "eval_runtime": 98.5702,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 1640
+ },
+ {
+ "entropy": 0.4952052327302786,
+ "epoch": 4.0,
+ "grad_norm": 2.080777168273926,
+ "learning_rate": 5.424453184004336e-05,
+ "loss": 0.4525449752807617,
+ "mean_token_accuracy": 0.8622837005517422,
+ "num_tokens": 4049116.0,
+ "step": 1660
+ },
+ {
+ "epoch": 4.0,
+ "eval_entropy": 0.5197472232446242,
+ "eval_loss": 0.57077956199646,
+ "eval_mean_token_accuracy": 0.8381094560864266,
+ "eval_num_tokens": 4049116.0,
+ "eval_runtime": 98.5498,
+ "eval_samples_per_second": 14.409,
+ "eval_steps_per_second": 1.806,
+ "step": 1660
+ },
+ {
+ "entropy": 0.42416453957557676,
+ "epoch": 4.048250904704463,
+ "grad_norm": 0.41175922751426697,
+ "learning_rate": 5.3715660220408146e-05,
+ "loss": 0.3610882043838501,
+ "mean_token_accuracy": 0.8832842499017716,
+ "num_tokens": 4097986.0,
+ "step": 1680
+ },
+ {
+ "epoch": 4.048250904704463,
+ "eval_entropy": 0.4814340365066957,
+ "eval_loss": 0.5963674783706665,
+ "eval_mean_token_accuracy": 0.8364584405770462,
+ "eval_num_tokens": 4097986.0,
+ "eval_runtime": 98.4464,
+ "eval_samples_per_second": 14.424,
+ "eval_steps_per_second": 1.808,
+ "step": 1680
+ },
+ {
+ "entropy": 0.4193890765309334,
+ "epoch": 4.0965018094089265,
+ "grad_norm": 0.4773254990577698,
+ "learning_rate": 5.3181816445467126e-05,
+ "loss": 0.36023797988891604,
+ "mean_token_accuracy": 0.884614959359169,
+ "num_tokens": 4147967.0,
+ "step": 1700
+ },
+ {
+ "epoch": 4.0965018094089265,
+ "eval_entropy": 0.47795870448096417,
+ "eval_loss": 0.6029345393180847,
+ "eval_mean_token_accuracy": 0.8354013039824668,
+ "eval_num_tokens": 4147967.0,
+ "eval_runtime": 98.5784,
+ "eval_samples_per_second": 14.405,
+ "eval_steps_per_second": 1.806,
+ "step": 1700
+ },
+ {
+ "entropy": 0.41848619729280473,
+ "epoch": 4.144752714113389,
+ "grad_norm": 0.7293462157249451,
+ "learning_rate": 5.264315158669384e-05,
+ "loss": 0.36621780395507814,
+ "mean_token_accuracy": 0.882377204298973,
+ "num_tokens": 4198623.0,
+ "step": 1720
+ },
+ {
+ "epoch": 4.144752714113389,
+ "eval_entropy": 0.4683974619661824,
+ "eval_loss": 0.6010014414787292,
+ "eval_mean_token_accuracy": 0.8358984790491254,
+ "eval_num_tokens": 4198623.0,
+ "eval_runtime": 98.5406,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 1720
+ },
+ {
+ "entropy": 0.41377718672156333,
+ "epoch": 4.193003618817853,
+ "grad_norm": 0.5061255693435669,
+ "learning_rate": 5.209981807987152e-05,
+ "loss": 0.36024787425994875,
+ "mean_token_accuracy": 0.8844522640109063,
+ "num_tokens": 4248406.0,
+ "step": 1740
+ },
+ {
+ "epoch": 4.193003618817853,
+ "eval_entropy": 0.4666026450609893,
+ "eval_loss": 0.6050319075584412,
+ "eval_mean_token_accuracy": 0.8357557922266843,
+ "eval_num_tokens": 4248406.0,
+ "eval_runtime": 98.6056,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 1740
+ },
+ {
+ "entropy": 0.4185848042368889,
+ "epoch": 4.241254523522316,
+ "grad_norm": 0.49589282274246216,
+ "learning_rate": 5.155196968195561e-05,
+ "loss": 0.36044313907623293,
+ "mean_token_accuracy": 0.8863146558403969,
+ "num_tokens": 4296608.0,
+ "step": 1760
+ },
+ {
+ "epoch": 4.241254523522316,
+ "eval_entropy": 0.4660757115382827,
+ "eval_loss": 0.6089843511581421,
+ "eval_mean_token_accuracy": 0.8352006492989786,
+ "eval_num_tokens": 4296608.0,
+ "eval_runtime": 98.5425,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 1760
+ },
+ {
+ "entropy": 0.4249664135277271,
+ "epoch": 4.2895054282267795,
+ "grad_norm": 0.6652432680130005,
+ "learning_rate": 5.099976142756227e-05,
+ "loss": 0.37266385555267334,
+ "mean_token_accuracy": 0.880594964325428,
+ "num_tokens": 4345389.0,
+ "step": 1780
+ },
+ {
+ "epoch": 4.2895054282267795,
+ "eval_entropy": 0.48157174483443915,
+ "eval_loss": 0.5987246036529541,
+ "eval_mean_token_accuracy": 0.8361060595914219,
+ "eval_num_tokens": 4345389.0,
+ "eval_runtime": 98.4652,
+ "eval_samples_per_second": 14.421,
+ "eval_steps_per_second": 1.808,
+ "step": 1780
+ },
+ {
+ "entropy": 0.4195803053677082,
+ "epoch": 4.337756332931242,
+ "grad_norm": 0.6217629313468933,
+ "learning_rate": 5.044334958509551e-05,
+ "loss": 0.3646178483963013,
+ "mean_token_accuracy": 0.8808894723653793,
+ "num_tokens": 4395878.0,
+ "step": 1800
+ },
+ {
+ "epoch": 4.337756332931242,
+ "eval_entropy": 0.47762840182593697,
+ "eval_loss": 0.5994009375572205,
+ "eval_mean_token_accuracy": 0.8366789422678144,
+ "eval_num_tokens": 4395878.0,
+ "eval_runtime": 98.6677,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 1800
+ },
+ {
+ "entropy": 0.42473565861582757,
+ "epoch": 4.386007237635706,
+ "grad_norm": 0.6135475635528564,
+ "learning_rate": 4.9882891612525023e-05,
+ "loss": 0.37249753475189207,
+ "mean_token_accuracy": 0.8800962314009666,
+ "num_tokens": 4440908.0,
+ "step": 1820
+ },
+ {
+ "epoch": 4.386007237635706,
+ "eval_entropy": 0.47359335924802204,
+ "eval_loss": 0.6024664640426636,
+ "eval_mean_token_accuracy": 0.8355274846714534,
+ "eval_num_tokens": 4440908.0,
+ "eval_runtime": 98.5103,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 1820
+ },
+ {
+ "entropy": 0.4212189674377441,
+ "epoch": 4.434258142340169,
+ "grad_norm": 0.5874977707862854,
+ "learning_rate": 4.931854611282748e-05,
+ "loss": 0.36766781806945803,
+ "mean_token_accuracy": 0.8837772741913795,
+ "num_tokens": 4489671.0,
+ "step": 1840
+ },
+ {
+ "epoch": 4.434258142340169,
+ "eval_entropy": 0.46670635798004234,
+ "eval_loss": 0.6085236668586731,
+ "eval_mean_token_accuracy": 0.8355892652206207,
+ "eval_num_tokens": 4489671.0,
+ "eval_runtime": 98.5871,
+ "eval_samples_per_second": 14.404,
+ "eval_steps_per_second": 1.806,
+ "step": 1840
+ },
+ {
+ "entropy": 0.41604730263352396,
+ "epoch": 4.4825090470446325,
+ "grad_norm": 0.5763360261917114,
+ "learning_rate": 4.875047278910371e-05,
+ "loss": 0.36643099784851074,
+ "mean_token_accuracy": 0.8825299561023712,
+ "num_tokens": 4541286.0,
+ "step": 1860
+ },
+ {
+ "epoch": 4.4825090470446325,
+ "eval_entropy": 0.47700903121005284,
+ "eval_loss": 0.606238842010498,
+ "eval_mean_token_accuracy": 0.8348092550642034,
+ "eval_num_tokens": 4541286.0,
+ "eval_runtime": 98.5305,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 1860
+ },
+ {
+ "entropy": 0.4189125746488571,
+ "epoch": 4.530759951749095,
+ "grad_norm": 0.5916087627410889,
+ "learning_rate": 4.817883239938471e-05,
+ "loss": 0.3703711748123169,
+ "mean_token_accuracy": 0.8825044095516205,
+ "num_tokens": 4589986.0,
+ "step": 1880
+ },
+ {
+ "epoch": 4.530759951749095,
+ "eval_entropy": 0.47982015703501324,
+ "eval_loss": 0.6003754138946533,
+ "eval_mean_token_accuracy": 0.8351969742373134,
+ "eval_num_tokens": 4589986.0,
+ "eval_runtime": 98.4111,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 1880
+ },
+ {
+ "entropy": 0.4429341182112694,
+ "epoch": 4.579010856453559,
+ "grad_norm": 0.6441562175750732,
+ "learning_rate": 4.7603786711138866e-05,
+ "loss": 0.3824815511703491,
+ "mean_token_accuracy": 0.8784194126725197,
+ "num_tokens": 4636548.0,
+ "step": 1900
+ },
+ {
+ "epoch": 4.579010856453559,
+ "eval_entropy": 0.4740966230296017,
+ "eval_loss": 0.5988267660140991,
+ "eval_mean_token_accuracy": 0.8369577316085944,
+ "eval_num_tokens": 4636548.0,
+ "eval_runtime": 98.5291,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 1900
+ },
+ {
+ "entropy": 0.41137003526091576,
+ "epoch": 4.627261761158022,
+ "grad_norm": 0.6663628816604614,
+ "learning_rate": 4.702549845549385e-05,
+ "loss": 0.36178064346313477,
+ "mean_token_accuracy": 0.8830229461193084,
+ "num_tokens": 4686918.0,
+ "step": 1920
+ },
+ {
+ "epoch": 4.627261761158022,
+ "eval_entropy": 0.4730259802904022,
+ "eval_loss": 0.6001402139663696,
+ "eval_mean_token_accuracy": 0.8366120139534554,
+ "eval_num_tokens": 4686918.0,
+ "eval_runtime": 98.5556,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 1920
+ },
+ {
+ "entropy": 0.4264552935957909,
+ "epoch": 4.675512665862485,
+ "grad_norm": 0.6708512902259827,
+ "learning_rate": 4.64441312811855e-05,
+ "loss": 0.376303505897522,
+ "mean_token_accuracy": 0.8826978862285614,
+ "num_tokens": 4734637.0,
+ "step": 1940
+ },
+ {
+ "epoch": 4.675512665862485,
+ "eval_entropy": 0.4687648815720269,
+ "eval_loss": 0.5982367992401123,
+ "eval_mean_token_accuracy": 0.8368918728292658,
+ "eval_num_tokens": 4734637.0,
+ "eval_runtime": 98.4928,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 1940
+ },
+ {
+ "entropy": 0.4378764644265175,
+ "epoch": 4.723763570566948,
+ "grad_norm": 0.7098956108093262,
+ "learning_rate": 4.585984970824725e-05,
+ "loss": 0.38264145851135256,
+ "mean_token_accuracy": 0.8768767058849335,
+ "num_tokens": 4783109.0,
+ "step": 1960
+ },
+ {
+ "epoch": 4.723763570566948,
+ "eval_entropy": 0.46603381248672354,
+ "eval_loss": 0.6010599732398987,
+ "eval_mean_token_accuracy": 0.8361037319965576,
+ "eval_num_tokens": 4783109.0,
+ "eval_runtime": 98.5157,
+ "eval_samples_per_second": 14.414,
+ "eval_steps_per_second": 1.807,
+ "step": 1960
+ },
+ {
+ "entropy": 0.44009693413972856,
+ "epoch": 4.772014475271411,
+ "grad_norm": 0.7529758214950562,
+ "learning_rate": 4.527281908145284e-05,
+ "loss": 0.379723048210144,
+ "mean_token_accuracy": 0.8799900904297828,
+ "num_tokens": 4828531.0,
+ "step": 1980
+ },
+ {
+ "epoch": 4.772014475271411,
+ "eval_entropy": 0.47042709873633437,
+ "eval_loss": 0.5984888672828674,
+ "eval_mean_token_accuracy": 0.8375813492228475,
+ "eval_num_tokens": 4828531.0,
+ "eval_runtime": 98.3352,
+ "eval_samples_per_second": 14.44,
+ "eval_steps_per_second": 1.81,
+ "step": 1980
+ },
+ {
+ "entropy": 0.43963938802480695,
+ "epoch": 4.820265379975875,
+ "grad_norm": 0.530246376991272,
+ "learning_rate": 4.468320552352577e-05,
+ "loss": 0.3855365037918091,
+ "mean_token_accuracy": 0.8766040325164794,
+ "num_tokens": 4877038.0,
+ "step": 2000
+ },
+ {
+ "epoch": 4.820265379975875,
+ "eval_entropy": 0.4773505919435051,
+ "eval_loss": 0.5956445932388306,
+ "eval_mean_token_accuracy": 0.8365653522228926,
+ "eval_num_tokens": 4877038.0,
+ "eval_runtime": 98.6015,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 2000
+ },
+ {
+ "entropy": 0.4241503834724426,
+ "epoch": 4.868516284680338,
+ "grad_norm": 0.5264645218849182,
+ "learning_rate": 4.409117588812845e-05,
+ "loss": 0.3725905418395996,
+ "mean_token_accuracy": 0.8816388875246048,
+ "num_tokens": 4928583.0,
+ "step": 2020
+ },
+ {
+ "epoch": 4.868516284680338,
+ "eval_entropy": 0.4738255136468437,
+ "eval_loss": 0.5987666845321655,
+ "eval_mean_token_accuracy": 0.8362959995028678,
+ "eval_num_tokens": 4928583.0,
+ "eval_runtime": 98.6141,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 2020
+ },
+ {
+ "entropy": 0.4451489016413689,
+ "epoch": 4.916767189384801,
+ "grad_norm": 0.5322316884994507,
+ "learning_rate": 4.3496897712644684e-05,
+ "loss": 0.3903519630432129,
+ "mean_token_accuracy": 0.8752380400896073,
+ "num_tokens": 4975223.0,
+ "step": 2040
+ },
+ {
+ "epoch": 4.916767189384801,
+ "eval_entropy": 0.4662560015247109,
+ "eval_loss": 0.5963740944862366,
+ "eval_mean_token_accuracy": 0.8378168083308788,
+ "eval_num_tokens": 4975223.0,
+ "eval_runtime": 98.6613,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 2040
+ },
+ {
+ "entropy": 0.43658976554870604,
+ "epoch": 4.965018094089264,
+ "grad_norm": 0.6989460587501526,
+ "learning_rate": 4.29005391707685e-05,
+ "loss": 0.3851702451705933,
+ "mean_token_accuracy": 0.8790383815765381,
+ "num_tokens": 5021142.0,
+ "step": 2060
+ },
+ {
+ "epoch": 4.965018094089264,
+ "eval_entropy": 0.4791082282749455,
+ "eval_loss": 0.5912667512893677,
+ "eval_mean_token_accuracy": 0.83739292286755,
+ "eval_num_tokens": 5021142.0,
+ "eval_runtime": 98.5606,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 2060
+ },
+ {
+ "entropy": 0.40598615316244274,
+ "epoch": 5.012062726176116,
+ "grad_norm": 0.5303810238838196,
+ "learning_rate": 4.2302269024913e-05,
+ "loss": 0.3450628280639648,
+ "mean_token_accuracy": 0.8893755567379487,
+ "num_tokens": 5073226.0,
+ "step": 2080
+ },
+ {
+ "epoch": 5.012062726176116,
+ "eval_entropy": 0.42836171187711564,
+ "eval_loss": 0.6335396766662598,
+ "eval_mean_token_accuracy": 0.8349329397249757,
+ "eval_num_tokens": 5073226.0,
+ "eval_runtime": 98.3503,
+ "eval_samples_per_second": 14.438,
+ "eval_steps_per_second": 1.81,
+ "step": 2080
+ },
+ {
+ "entropy": 0.3517624609172344,
+ "epoch": 5.060313630880579,
+ "grad_norm": 0.6402326226234436,
+ "learning_rate": 4.170225657845261e-05,
+ "loss": 0.2898552894592285,
+ "mean_token_accuracy": 0.904600378870964,
+ "num_tokens": 5120052.0,
+ "step": 2100
+ },
+ {
+ "epoch": 5.060313630880579,
+ "eval_entropy": 0.4094204810563098,
+ "eval_loss": 0.652080237865448,
+ "eval_mean_token_accuracy": 0.8331165705504042,
+ "eval_num_tokens": 5120052.0,
+ "eval_runtime": 98.3573,
+ "eval_samples_per_second": 14.437,
+ "eval_steps_per_second": 1.81,
+ "step": 2100
+ },
+ {
+ "entropy": 0.3560003057122231,
+ "epoch": 5.108564535585042,
+ "grad_norm": 0.9274800419807434,
+ "learning_rate": 4.110067162781221e-05,
+ "loss": 0.2926265954971313,
+ "mean_token_accuracy": 0.9040415585041046,
+ "num_tokens": 5165984.0,
+ "step": 2120
+ },
+ {
+ "epoch": 5.108564535585042,
+ "eval_entropy": 0.41318756994906436,
+ "eval_loss": 0.6567395925521851,
+ "eval_mean_token_accuracy": 0.83362964197491,
+ "eval_num_tokens": 5165984.0,
+ "eval_runtime": 98.5042,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 2120
+ },
+ {
+ "entropy": 0.35080533251166346,
+ "epoch": 5.156815440289505,
+ "grad_norm": 0.6713427901268005,
+ "learning_rate": 4.049768441441677e-05,
+ "loss": 0.29822506904602053,
+ "mean_token_accuracy": 0.9027505293488503,
+ "num_tokens": 5214442.0,
+ "step": 2140
+ },
+ {
+ "epoch": 5.156815440289505,
+ "eval_entropy": 0.416848015416874,
+ "eval_loss": 0.6563553214073181,
+ "eval_mean_token_accuracy": 0.8333374677079447,
+ "eval_num_tokens": 5214442.0,
+ "eval_runtime": 98.4261,
+ "eval_samples_per_second": 14.427,
+ "eval_steps_per_second": 1.808,
+ "step": 2140
+ },
+ {
+ "entropy": 0.3736354000866413,
+ "epoch": 5.205066344993969,
+ "grad_norm": 0.5667551755905151,
+ "learning_rate": 3.989346557651503e-05,
+ "loss": 0.3088909387588501,
+ "mean_token_accuracy": 0.9012129113078118,
+ "num_tokens": 5260853.0,
+ "step": 2160
+ },
+ {
+ "epoch": 5.205066344993969,
+ "eval_entropy": 0.41974830309326727,
+ "eval_loss": 0.6504881381988525,
+ "eval_mean_token_accuracy": 0.8349264759695931,
+ "eval_num_tokens": 5260853.0,
+ "eval_runtime": 98.5007,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 2160
+ },
+ {
+ "entropy": 0.35063633099198344,
+ "epoch": 5.253317249698432,
+ "grad_norm": 0.7737690210342407,
+ "learning_rate": 3.9288186100890903e-05,
+ "loss": 0.2888066053390503,
+ "mean_token_accuracy": 0.9040960729122162,
+ "num_tokens": 5312338.0,
+ "step": 2180
+ },
+ {
+ "epoch": 5.253317249698432,
+ "eval_entropy": 0.4058936465992017,
+ "eval_loss": 0.6578952670097351,
+ "eval_mean_token_accuracy": 0.8352396407823884,
+ "eval_num_tokens": 5312338.0,
+ "eval_runtime": 98.4807,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 2180
+ },
+ {
+ "entropy": 0.3545838728547096,
+ "epoch": 5.301568154402895,
+ "grad_norm": 0.8293929100036621,
+ "learning_rate": 3.8682017274476215e-05,
+ "loss": 0.29347708225250246,
+ "mean_token_accuracy": 0.902638104557991,
+ "num_tokens": 5359039.0,
+ "step": 2200
+ },
+ {
+ "epoch": 5.301568154402895,
+ "eval_entropy": 0.41738414345832353,
+ "eval_loss": 0.6543260812759399,
+ "eval_mean_token_accuracy": 0.8336448260907376,
+ "eval_num_tokens": 5359039.0,
+ "eval_runtime": 98.499,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 2200
+ },
+ {
+ "entropy": 0.3416903533041477,
+ "epoch": 5.349819059107358,
+ "grad_norm": 0.7580088973045349,
+ "learning_rate": 3.8075130635878556e-05,
+ "loss": 0.28760058879852296,
+ "mean_token_accuracy": 0.906610831618309,
+ "num_tokens": 5409807.0,
+ "step": 2220
+ },
+ {
+ "epoch": 5.349819059107358,
+ "eval_entropy": 0.4121073748958245,
+ "eval_loss": 0.6563934087753296,
+ "eval_mean_token_accuracy": 0.833708616120092,
+ "eval_num_tokens": 5409807.0,
+ "eval_runtime": 98.3808,
+ "eval_samples_per_second": 14.434,
+ "eval_steps_per_second": 1.809,
+ "step": 2220
+ },
+ {
+ "entropy": 0.352034630253911,
+ "epoch": 5.398069963811821,
+ "grad_norm": 0.7154259085655212,
+ "learning_rate": 3.746769792683787e-05,
+ "loss": 0.2967940092086792,
+ "mean_token_accuracy": 0.9023159235715866,
+ "num_tokens": 5458308.0,
+ "step": 2240
+ },
+ {
+ "epoch": 5.398069963811821,
+ "eval_entropy": 0.4149914874454563,
+ "eval_loss": 0.6444063186645508,
+ "eval_mean_token_accuracy": 0.8360218017288809,
+ "eval_num_tokens": 5458308.0,
+ "eval_runtime": 98.4832,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 2240
+ },
+ {
+ "entropy": 0.3539439059793949,
+ "epoch": 5.446320868516285,
+ "grad_norm": 0.8078773617744446,
+ "learning_rate": 3.685989104362558e-05,
+ "loss": 0.2928986310958862,
+ "mean_token_accuracy": 0.905618867278099,
+ "num_tokens": 5505380.0,
+ "step": 2260
+ },
+ {
+ "epoch": 5.446320868516285,
+ "eval_entropy": 0.4099769044793054,
+ "eval_loss": 0.6588519215583801,
+ "eval_mean_token_accuracy": 0.8353862765799748,
+ "eval_num_tokens": 5505380.0,
+ "eval_runtime": 98.3781,
+ "eval_samples_per_second": 14.434,
+ "eval_steps_per_second": 1.809,
+ "step": 2260
+ },
+ {
+ "entropy": 0.35821759328246117,
+ "epoch": 5.4945717732207475,
+ "grad_norm": 0.8760809302330017,
+ "learning_rate": 3.6251881988399994e-05,
+ "loss": 0.29706525802612305,
+ "mean_token_accuracy": 0.9018172711133957,
+ "num_tokens": 5553385.0,
+ "step": 2280
+ },
+ {
+ "epoch": 5.4945717732207475,
+ "eval_entropy": 0.40791591667057425,
+ "eval_loss": 0.6583756804466248,
+ "eval_mean_token_accuracy": 0.8349186125096311,
+ "eval_num_tokens": 5553385.0,
+ "eval_runtime": 98.4808,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 2280
+ },
+ {
+ "entropy": 0.34672123566269875,
+ "epoch": 5.542822677925211,
+ "grad_norm": 0.8768810033798218,
+ "learning_rate": 3.564384282053165e-05,
+ "loss": 0.2885677576065063,
+ "mean_token_accuracy": 0.9063566774129868,
+ "num_tokens": 5604413.0,
+ "step": 2300
+ },
+ {
+ "epoch": 5.542822677925211,
+ "eval_entropy": 0.412923522377282,
+ "eval_loss": 0.6535082459449768,
+ "eval_mean_token_accuracy": 0.8350685978873392,
+ "eval_num_tokens": 5604413.0,
+ "eval_runtime": 98.3842,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 2300
+ },
+ {
+ "entropy": 0.35606732666492463,
+ "epoch": 5.591073582629674,
+ "grad_norm": 0.7237412929534912,
+ "learning_rate": 3.503594560791267e-05,
+ "loss": 0.30418403148651124,
+ "mean_token_accuracy": 0.9004160806536674,
+ "num_tokens": 5657218.0,
+ "step": 2320
+ },
+ {
+ "epoch": 5.591073582629674,
+ "eval_entropy": 0.42160589155856143,
+ "eval_loss": 0.6412338018417358,
+ "eval_mean_token_accuracy": 0.834980196162556,
+ "eval_num_tokens": 5657218.0,
+ "eval_runtime": 98.3674,
+ "eval_samples_per_second": 14.436,
+ "eval_steps_per_second": 1.81,
+ "step": 2320
+ },
+ {
+ "entropy": 0.35330103561282156,
+ "epoch": 5.639324487334138,
+ "grad_norm": 0.5864736437797546,
+ "learning_rate": 3.442836237826349e-05,
+ "loss": 0.29608047008514404,
+ "mean_token_accuracy": 0.9029312193393707,
+ "num_tokens": 5709140.0,
+ "step": 2340
+ },
+ {
+ "epoch": 5.639324487334138,
+ "eval_entropy": 0.4273778943533308,
+ "eval_loss": 0.6407549977302551,
+ "eval_mean_token_accuracy": 0.8351332397943132,
+ "eval_num_tokens": 5709140.0,
+ "eval_runtime": 98.5533,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 2340
+ },
+ {
+ "entropy": 0.353700140863657,
+ "epoch": 5.6875753920386005,
+ "grad_norm": 0.744305431842804,
+ "learning_rate": 3.382126507045109e-05,
+ "loss": 0.3004793643951416,
+ "mean_token_accuracy": 0.9015571728348732,
+ "num_tokens": 5760973.0,
+ "step": 2360
+ },
+ {
+ "epoch": 5.6875753920386005,
+ "eval_entropy": 0.4262272163053577,
+ "eval_loss": 0.6443405151367188,
+ "eval_mean_token_accuracy": 0.8354629797212193,
+ "eval_num_tokens": 5760973.0,
+ "eval_runtime": 98.3719,
+ "eval_samples_per_second": 14.435,
+ "eval_steps_per_second": 1.809,
+ "step": 2360
+ },
+ {
+ "entropy": 0.36056690439581873,
+ "epoch": 5.735826296743064,
+ "grad_norm": 0.6552110910415649,
+ "learning_rate": 3.321482548583228e-05,
+ "loss": 0.2950709104537964,
+ "mean_token_accuracy": 0.9033273816108703,
+ "num_tokens": 5808058.0,
+ "step": 2380
+ },
+ {
+ "epoch": 5.735826296743064,
+ "eval_entropy": 0.41902332275771026,
+ "eval_loss": 0.6472867727279663,
+ "eval_mean_token_accuracy": 0.8354150207524889,
+ "eval_num_tokens": 5808058.0,
+ "eval_runtime": 98.3956,
+ "eval_samples_per_second": 14.432,
+ "eval_steps_per_second": 1.809,
+ "step": 2380
+ },
+ {
+ "entropy": 0.35097662582993505,
+ "epoch": 5.784077201447527,
+ "grad_norm": 0.670831024646759,
+ "learning_rate": 3.260921523963606e-05,
+ "loss": 0.29224560260772703,
+ "mean_token_accuracy": 0.9042141750454903,
+ "num_tokens": 5857700.0,
+ "step": 2400
+ },
+ {
+ "epoch": 5.784077201447527,
+ "eval_entropy": 0.4284254325909561,
+ "eval_loss": 0.6407246589660645,
+ "eval_mean_token_accuracy": 0.8345498831084605,
+ "eval_num_tokens": 5857700.0,
+ "eval_runtime": 98.3651,
+ "eval_samples_per_second": 14.436,
+ "eval_steps_per_second": 1.81,
+ "step": 2400
+ },
+ {
+ "entropy": 0.35357567593455314,
+ "epoch": 5.832328106151991,
+ "grad_norm": 0.8287181258201599,
+ "learning_rate": 3.200460571239839e-05,
+ "loss": 0.2948127746582031,
+ "mean_token_accuracy": 0.9042218685150146,
+ "num_tokens": 5907064.0,
+ "step": 2420
+ },
+ {
+ "epoch": 5.832328106151991,
+ "eval_entropy": 0.4094803942723221,
+ "eval_loss": 0.6497060060501099,
+ "eval_mean_token_accuracy": 0.8358801147910986,
+ "eval_num_tokens": 5907064.0,
+ "eval_runtime": 98.3681,
+ "eval_samples_per_second": 14.436,
+ "eval_steps_per_second": 1.81,
+ "step": 2420
+ },
+ {
+ "entropy": 0.36021770909428596,
+ "epoch": 5.8805790108564535,
+ "grad_norm": 0.6800532937049866,
+ "learning_rate": 3.140116800146367e-05,
+ "loss": 0.30136709213256835,
+ "mean_token_accuracy": 0.9020729199051857,
+ "num_tokens": 5954440.0,
+ "step": 2440
+ },
+ {
+ "epoch": 5.8805790108564535,
+ "eval_entropy": 0.4202987858083811,
+ "eval_loss": 0.6428810954093933,
+ "eval_mean_token_accuracy": 0.8355459637856215,
+ "eval_num_tokens": 5954440.0,
+ "eval_runtime": 98.5147,
+ "eval_samples_per_second": 14.414,
+ "eval_steps_per_second": 1.807,
+ "step": 2440
+ },
+ {
+ "entropy": 0.34706664383411406,
+ "epoch": 5.928829915560916,
+ "grad_norm": 0.6911236643791199,
+ "learning_rate": 3.07990728725661e-05,
+ "loss": 0.28953113555908205,
+ "mean_token_accuracy": 0.9055185422301293,
+ "num_tokens": 6002089.0,
+ "step": 2460
+ },
+ {
+ "epoch": 5.928829915560916,
+ "eval_entropy": 0.406153663825453,
+ "eval_loss": 0.6541930437088013,
+ "eval_mean_token_accuracy": 0.834371502814668,
+ "eval_num_tokens": 6002089.0,
+ "eval_runtime": 98.3494,
+ "eval_samples_per_second": 14.438,
+ "eval_steps_per_second": 1.81,
+ "step": 2460
+ },
+ {
+ "entropy": 0.34852043315768244,
+ "epoch": 5.97708082026538,
+ "grad_norm": 0.7544664144515991,
+ "learning_rate": 3.0198490711505038e-05,
+ "loss": 0.28950369358062744,
+ "mean_token_accuracy": 0.9057677388191223,
+ "num_tokens": 6051074.0,
+ "step": 2480
+ },
+ {
+ "epoch": 5.97708082026538,
+ "eval_entropy": 0.4174931427735961,
+ "eval_loss": 0.6448875665664673,
+ "eval_mean_token_accuracy": 0.8358313615402478,
+ "eval_num_tokens": 6051074.0,
+ "eval_runtime": 98.4786,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 2480
+ },
+ {
+ "entropy": 0.31319292042499935,
+ "epoch": 6.024125452352232,
+ "grad_norm": 0.7616425156593323,
+ "learning_rate": 2.9599591475927893e-05,
+ "loss": 0.24788825511932372,
+ "mean_token_accuracy": 0.9198053983541635,
+ "num_tokens": 6101526.0,
+ "step": 2500
+ },
+ {
+ "epoch": 6.024125452352232,
+ "eval_entropy": 0.3601247634780541,
+ "eval_loss": 0.7335909008979797,
+ "eval_mean_token_accuracy": 0.8309332149082356,
+ "eval_num_tokens": 6101526.0,
+ "eval_runtime": 98.3569,
+ "eval_samples_per_second": 14.437,
+ "eval_steps_per_second": 1.81,
+ "step": 2500
+ },
+ {
+ "entropy": 0.27905523739755156,
+ "epoch": 6.072376357056695,
+ "grad_norm": 0.7366685271263123,
+ "learning_rate": 2.9002544647234015e-05,
+ "loss": 0.21279211044311525,
+ "mean_token_accuracy": 0.929847776889801,
+ "num_tokens": 6147939.0,
+ "step": 2520
+ },
+ {
+ "epoch": 6.072376357056695,
+ "eval_entropy": 0.36974305969275784,
+ "eval_loss": 0.7080109715461731,
+ "eval_mean_token_accuracy": 0.8330447111236915,
+ "eval_num_tokens": 6147939.0,
+ "eval_runtime": 98.4908,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 2520
+ },
+ {
+ "entropy": 0.2782863024622202,
+ "epoch": 6.120627261761158,
+ "grad_norm": 0.9167613387107849,
+ "learning_rate": 2.840751918261356e-05,
+ "loss": 0.20897085666656495,
+ "mean_token_accuracy": 0.9294204145669938,
+ "num_tokens": 6193631.0,
+ "step": 2540
+ },
+ {
+ "epoch": 6.120627261761158,
+ "eval_entropy": 0.366466105319141,
+ "eval_loss": 0.7170661091804504,
+ "eval_mean_token_accuracy": 0.8328661680891273,
+ "eval_num_tokens": 6193631.0,
+ "eval_runtime": 98.4385,
+ "eval_samples_per_second": 14.425,
+ "eval_steps_per_second": 1.808,
+ "step": 2540
+ },
+ {
+ "entropy": 0.28355731777846815,
+ "epoch": 6.168878166465621,
+ "grad_norm": 0.8010057806968689,
+ "learning_rate": 2.7814683467234538e-05,
+ "loss": 0.211204195022583,
+ "mean_token_accuracy": 0.9297966167330742,
+ "num_tokens": 6240258.0,
+ "step": 2560
+ },
+ {
+ "epoch": 6.168878166465621,
+ "eval_entropy": 0.36755596988656547,
+ "eval_loss": 0.7274795174598694,
+ "eval_mean_token_accuracy": 0.8303602230682802,
+ "eval_num_tokens": 6240258.0,
+ "eval_runtime": 98.487,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 2560
+ },
+ {
+ "entropy": 0.27822012901306153,
+ "epoch": 6.217129071170085,
+ "grad_norm": 0.7516391277313232,
+ "learning_rate": 2.7224205266591812e-05,
+ "loss": 0.2117089033126831,
+ "mean_token_accuracy": 0.9294167309999466,
+ "num_tokens": 6289137.0,
+ "step": 2580
+ },
+ {
+ "epoch": 6.217129071170085,
+ "eval_entropy": 0.37035224410925016,
+ "eval_loss": 0.7123831510543823,
+ "eval_mean_token_accuracy": 0.8318134563022785,
+ "eval_num_tokens": 6289137.0,
+ "eval_runtime": 98.5029,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 2580
+ },
+ {
+ "entropy": 0.2784307342022657,
+ "epoch": 6.265379975874548,
+ "grad_norm": 0.8481637239456177,
+ "learning_rate": 2.6636251679031508e-05,
+ "loss": 0.21476459503173828,
+ "mean_token_accuracy": 0.9293668136000633,
+ "num_tokens": 6339215.0,
+ "step": 2600
+ },
+ {
+ "epoch": 6.265379975874548,
+ "eval_entropy": 0.3675622491354353,
+ "eval_loss": 0.7221675515174866,
+ "eval_mean_token_accuracy": 0.8314910557832611,
+ "eval_num_tokens": 6339215.0,
+ "eval_runtime": 98.3892,
+ "eval_samples_per_second": 14.432,
+ "eval_steps_per_second": 1.809,
+ "step": 2600
+ },
+ {
+ "entropy": 0.27380565330386164,
+ "epoch": 6.3136308805790105,
+ "grad_norm": 0.7563941478729248,
+ "learning_rate": 2.605098908846406e-05,
+ "loss": 0.2117525815963745,
+ "mean_token_accuracy": 0.9299218639731407,
+ "num_tokens": 6390057.0,
+ "step": 2620
+ },
+ {
+ "epoch": 6.3136308805790105,
+ "eval_entropy": 0.3676499978879864,
+ "eval_loss": 0.7249981164932251,
+ "eval_mean_token_accuracy": 0.8315788115678209,
+ "eval_num_tokens": 6390057.0,
+ "eval_runtime": 98.491,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 2620
+ },
+ {
+ "entropy": 0.26819119676947595,
+ "epoch": 6.361881785283474,
+ "grad_norm": 0.9161278605461121,
+ "learning_rate": 2.546858311727963e-05,
+ "loss": 0.2020425319671631,
+ "mean_token_accuracy": 0.9322459831833839,
+ "num_tokens": 6443300.0,
+ "step": 2640
+ },
+ {
+ "epoch": 6.361881785283474,
+ "eval_entropy": 0.3589264298422953,
+ "eval_loss": 0.737293004989624,
+ "eval_mean_token_accuracy": 0.8307386482699534,
+ "eval_num_tokens": 6443300.0,
+ "eval_runtime": 98.4767,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 2640
+ },
+ {
+ "entropy": 0.275980044901371,
+ "epoch": 6.410132689987937,
+ "grad_norm": 0.7255368828773499,
+ "learning_rate": 2.488919857947885e-05,
+ "loss": 0.21207733154296876,
+ "mean_token_accuracy": 0.9280627354979515,
+ "num_tokens": 6491445.0,
+ "step": 2660
+ },
+ {
+ "epoch": 6.410132689987937,
+ "eval_entropy": 0.3729423410102223,
+ "eval_loss": 0.7137707471847534,
+ "eval_mean_token_accuracy": 0.8315478007445175,
+ "eval_num_tokens": 6491445.0,
+ "eval_runtime": 98.384,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 2660
+ },
+ {
+ "entropy": 0.2794003043323755,
+ "epoch": 6.458383594692401,
+ "grad_norm": 0.817340612411499,
+ "learning_rate": 2.4312999434032465e-05,
+ "loss": 0.2155231714248657,
+ "mean_token_accuracy": 0.930096672475338,
+ "num_tokens": 6541352.0,
+ "step": 2680
+ },
+ {
+ "epoch": 6.458383594692401,
+ "eval_entropy": 0.36857131709543506,
+ "eval_loss": 0.7273393869400024,
+ "eval_mean_token_accuracy": 0.8302122903673836,
+ "eval_num_tokens": 6541352.0,
+ "eval_runtime": 98.5119,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 2680
+ },
+ {
+ "entropy": 0.28550401143729687,
+ "epoch": 6.5066344993968634,
+ "grad_norm": 0.7945013642311096,
+ "learning_rate": 2.374014873848286e-05,
+ "loss": 0.21428213119506836,
+ "mean_token_accuracy": 0.9295192703604698,
+ "num_tokens": 6589480.0,
+ "step": 2700
+ },
+ {
+ "epoch": 6.5066344993968634,
+ "eval_entropy": 0.3720237272843886,
+ "eval_loss": 0.7190349698066711,
+ "eval_mean_token_accuracy": 0.8307539365264807,
+ "eval_num_tokens": 6589480.0,
+ "eval_runtime": 98.3972,
+ "eval_samples_per_second": 14.431,
+ "eval_steps_per_second": 1.809,
+ "step": 2700
+ },
+ {
+ "entropy": 0.2789166793227196,
+ "epoch": 6.554885404101327,
+ "grad_norm": 0.7820216417312622,
+ "learning_rate": 2.3170808602800694e-05,
+ "loss": 0.21065421104431153,
+ "mean_token_accuracy": 0.9293019056320191,
+ "num_tokens": 6635214.0,
+ "step": 2720
+ },
+ {
+ "epoch": 6.554885404101327,
+ "eval_entropy": 0.36198576715555086,
+ "eval_loss": 0.7250529527664185,
+ "eval_mean_token_accuracy": 0.8310985079642093,
+ "eval_num_tokens": 6635214.0,
+ "eval_runtime": 98.516,
+ "eval_samples_per_second": 14.414,
+ "eval_steps_per_second": 1.807,
+ "step": 2720
+ },
+ {
+ "entropy": 0.2805891681462526,
+ "epoch": 6.60313630880579,
+ "grad_norm": 0.8589067459106445,
+ "learning_rate": 2.2605140143509734e-05,
+ "loss": 0.21532771587371827,
+ "mean_token_accuracy": 0.9301707550883294,
+ "num_tokens": 6682916.0,
+ "step": 2740
+ },
+ {
+ "epoch": 6.60313630880579,
+ "eval_entropy": 0.36839111149311066,
+ "eval_loss": 0.720038890838623,
+ "eval_mean_token_accuracy": 0.8314180384191234,
+ "eval_num_tokens": 6682916.0,
+ "eval_runtime": 98.3864,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 2740
+ },
+ {
+ "entropy": 0.2923452850431204,
+ "epoch": 6.651387213510254,
+ "grad_norm": 1.1464570760726929,
+ "learning_rate": 2.2043303438092746e-05,
+ "loss": 0.22156639099121095,
+ "mean_token_accuracy": 0.9259255439043045,
+ "num_tokens": 6728644.0,
+ "step": 2760
+ },
+ {
+ "epoch": 6.651387213510254,
+ "eval_entropy": 0.36308020152402726,
+ "eval_loss": 0.7283808588981628,
+ "eval_mean_token_accuracy": 0.8307237183110098,
+ "eval_num_tokens": 6728644.0,
+ "eval_runtime": 98.3733,
+ "eval_samples_per_second": 14.435,
+ "eval_steps_per_second": 1.809,
+ "step": 2760
+ },
+ {
+ "entropy": 0.2749974336475134,
+ "epoch": 6.699638118214716,
+ "grad_norm": 1.0338026285171509,
+ "learning_rate": 2.1485457479691453e-05,
+ "loss": 0.2111821413040161,
+ "mean_token_accuracy": 0.9308380514383316,
+ "num_tokens": 6776170.0,
+ "step": 2780
+ },
+ {
+ "epoch": 6.699638118214716,
+ "eval_entropy": 0.35679193979568696,
+ "eval_loss": 0.7316031455993652,
+ "eval_mean_token_accuracy": 0.8316076066386834,
+ "eval_num_tokens": 6776170.0,
+ "eval_runtime": 98.4012,
+ "eval_samples_per_second": 14.431,
+ "eval_steps_per_second": 1.809,
+ "step": 2780
+ },
+ {
+ "entropy": 0.2827217370271683,
+ "epoch": 6.74788902291918,
+ "grad_norm": 0.6860962510108948,
+ "learning_rate": 2.0931760132113435e-05,
+ "loss": 0.219130802154541,
+ "mean_token_accuracy": 0.9272046208381652,
+ "num_tokens": 6823320.0,
+ "step": 2800
+ },
+ {
+ "epoch": 6.74788902291918,
+ "eval_entropy": 0.367448085479522,
+ "eval_loss": 0.7210394144058228,
+ "eval_mean_token_accuracy": 0.8324049028118005,
+ "eval_num_tokens": 6823320.0,
+ "eval_runtime": 98.3867,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 2800
+ },
+ {
+ "entropy": 0.28139964006841184,
+ "epoch": 6.796139927623643,
+ "grad_norm": 0.7198498249053955,
+ "learning_rate": 2.038236808515841e-05,
+ "loss": 0.2179172992706299,
+ "mean_token_accuracy": 0.9282332047820091,
+ "num_tokens": 6874282.0,
+ "step": 2820
+ },
+ {
+ "epoch": 6.796139927623643,
+ "eval_entropy": 0.3658287344353922,
+ "eval_loss": 0.7204482555389404,
+ "eval_mean_token_accuracy": 0.8324238242727987,
+ "eval_num_tokens": 6874282.0,
+ "eval_runtime": 98.3796,
+ "eval_samples_per_second": 14.434,
+ "eval_steps_per_second": 1.809,
+ "step": 2820
+ },
+ {
+ "entropy": 0.28035391345620153,
+ "epoch": 6.844390832328106,
+ "grad_norm": 0.8470256924629211,
+ "learning_rate": 1.9837436810277026e-05,
+ "loss": 0.21534478664398193,
+ "mean_token_accuracy": 0.9284071952104569,
+ "num_tokens": 6923379.0,
+ "step": 2840
+ },
+ {
+ "epoch": 6.844390832328106,
+ "eval_entropy": 0.3615414013018769,
+ "eval_loss": 0.7223248481750488,
+ "eval_mean_token_accuracy": 0.8328172581249409,
+ "eval_num_tokens": 6923379.0,
+ "eval_runtime": 98.4426,
+ "eval_samples_per_second": 14.425,
+ "eval_steps_per_second": 1.808,
+ "step": 2840
+ },
+ {
+ "entropy": 0.27447608187794686,
+ "epoch": 6.892641737032569,
+ "grad_norm": 1.1274205446243286,
+ "learning_rate": 1.9297120516574132e-05,
+ "loss": 0.2151768684387207,
+ "mean_token_accuracy": 0.9284873619675637,
+ "num_tokens": 6971734.0,
+ "step": 2860
+ },
+ {
+ "epoch": 6.892641737032569,
+ "eval_entropy": 0.36038200892089456,
+ "eval_loss": 0.7199941873550415,
+ "eval_mean_token_accuracy": 0.8334562922461649,
+ "eval_num_tokens": 6971734.0,
+ "eval_runtime": 98.4807,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 2860
+ },
+ {
+ "entropy": 0.27447244003415106,
+ "epoch": 6.940892641737032,
+ "grad_norm": 0.857891857624054,
+ "learning_rate": 1.876157210716952e-05,
+ "loss": 0.21279585361480713,
+ "mean_token_accuracy": 0.9296008720993996,
+ "num_tokens": 7024380.0,
+ "step": 2880
+ },
+ {
+ "epoch": 6.940892641737032,
+ "eval_entropy": 0.3578391455197602,
+ "eval_loss": 0.7355349063873291,
+ "eval_mean_token_accuracy": 0.8319483108065101,
+ "eval_num_tokens": 7024380.0,
+ "eval_runtime": 98.3412,
+ "eval_samples_per_second": 14.44,
+ "eval_steps_per_second": 1.81,
+ "step": 2880
+ },
+ {
+ "entropy": 0.27788355611264703,
+ "epoch": 6.989143546441496,
+ "grad_norm": 0.8024702668190002,
+ "learning_rate": 1.8230943135928133e-05,
+ "loss": 0.21420254707336425,
+ "mean_token_accuracy": 0.9291928946971894,
+ "num_tokens": 7074034.0,
+ "step": 2900
+ },
+ {
+ "epoch": 6.989143546441496,
+ "eval_entropy": 0.35886382102296593,
+ "eval_loss": 0.7313408851623535,
+ "eval_mean_token_accuracy": 0.8322128275137269,
+ "eval_num_tokens": 7074034.0,
+ "eval_runtime": 98.4564,
+ "eval_samples_per_second": 14.423,
+ "eval_steps_per_second": 1.808,
+ "step": 2900
+ },
+ {
+ "entropy": 0.24298458527296019,
+ "epoch": 7.036188178528348,
+ "grad_norm": 0.8567366600036621,
+ "learning_rate": 1.7705383764572093e-05,
+ "loss": 0.15583932399749756,
+ "mean_token_accuracy": 0.9481525069628006,
+ "num_tokens": 7122753.0,
+ "step": 2920
+ },
+ {
+ "epoch": 7.036188178528348,
+ "eval_entropy": 0.3236005037520709,
+ "eval_loss": 0.8240236043930054,
+ "eval_mean_token_accuracy": 0.8270553162928378,
+ "eval_num_tokens": 7122753.0,
+ "eval_runtime": 98.3819,
+ "eval_samples_per_second": 14.434,
+ "eval_steps_per_second": 1.809,
+ "step": 2920
+ },
+ {
+ "entropy": 0.2195398263633251,
+ "epoch": 7.084439083232811,
+ "grad_norm": 0.8163788914680481,
+ "learning_rate": 1.7185042720186728e-05,
+ "loss": 0.14468494653701783,
+ "mean_token_accuracy": 0.9525706857442856,
+ "num_tokens": 7171505.0,
+ "step": 2940
+ },
+ {
+ "epoch": 7.084439083232811,
+ "eval_entropy": 0.32686926424503326,
+ "eval_loss": 0.8029564023017883,
+ "eval_mean_token_accuracy": 0.8280681498934713,
+ "eval_num_tokens": 7171505.0,
+ "eval_runtime": 98.7239,
+ "eval_samples_per_second": 14.384,
+ "eval_steps_per_second": 1.803,
+ "step": 2940
+ },
+ {
+ "entropy": 0.21137732304632664,
+ "epoch": 7.132689987937274,
+ "grad_norm": 0.7098442316055298,
+ "learning_rate": 1.6670067253132565e-05,
+ "loss": 0.1382979393005371,
+ "mean_token_accuracy": 0.9557823866605759,
+ "num_tokens": 7219913.0,
+ "step": 2960
+ },
+ {
+ "epoch": 7.132689987937274,
+ "eval_entropy": 0.32376345288887454,
+ "eval_loss": 0.8175393342971802,
+ "eval_mean_token_accuracy": 0.8281539598207796,
+ "eval_num_tokens": 7219913.0,
+ "eval_runtime": 98.4728,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 2960
+ },
+ {
+ "entropy": 0.21736963428556919,
+ "epoch": 7.180940892641737,
+ "grad_norm": 0.8196840286254883,
+ "learning_rate": 1.616060309537518e-05,
+ "loss": 0.14342793226242065,
+ "mean_token_accuracy": 0.9531146377325058,
+ "num_tokens": 7265991.0,
+ "step": 2980
+ },
+ {
+ "epoch": 7.180940892641737,
+ "eval_entropy": 0.3240856607810835,
+ "eval_loss": 0.816273033618927,
+ "eval_mean_token_accuracy": 0.8283853041991759,
+ "eval_num_tokens": 7265991.0,
+ "eval_runtime": 98.4989,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 2980
+ },
+ {
+ "entropy": 0.2128997717052698,
+ "epoch": 7.2291917973462,
+ "grad_norm": 0.7581701278686523,
+ "learning_rate": 1.5656794419244872e-05,
+ "loss": 0.14148796796798707,
+ "mean_token_accuracy": 0.9551757454872132,
+ "num_tokens": 7317184.0,
+ "step": 3000
+ },
+ {
+ "epoch": 7.2291917973462,
+ "eval_entropy": 0.3229605887043342,
+ "eval_loss": 0.8140764832496643,
+ "eval_mean_token_accuracy": 0.8288295952121864,
+ "eval_num_tokens": 7317184.0,
+ "eval_runtime": 98.4981,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 3000
+ },
+ {
+ "entropy": 0.214377661049366,
+ "epoch": 7.277442702050664,
+ "grad_norm": 0.6986305117607117,
+ "learning_rate": 1.5158783796637473e-05,
+ "loss": 0.14369987249374389,
+ "mean_token_accuracy": 0.9528444185853004,
+ "num_tokens": 7365364.0,
+ "step": 3020
+ },
+ {
+ "epoch": 7.277442702050664,
+ "eval_entropy": 0.32001526101251665,
+ "eval_loss": 0.8190910816192627,
+ "eval_mean_token_accuracy": 0.829054100459881,
+ "eval_num_tokens": 7365364.0,
+ "eval_runtime": 98.5783,
+ "eval_samples_per_second": 14.405,
+ "eval_steps_per_second": 1.806,
+ "step": 3020
+ },
+ {
+ "entropy": 0.20811213366687298,
+ "epoch": 7.325693606755126,
+ "grad_norm": 0.9613916277885437,
+ "learning_rate": 1.4666712158668313e-05,
+ "loss": 0.14105329513549805,
+ "mean_token_accuracy": 0.9532626882195473,
+ "num_tokens": 7415795.0,
+ "step": 3040
+ },
+ {
+ "epoch": 7.325693606755126,
+ "eval_entropy": 0.31941057892327895,
+ "eval_loss": 0.8219555616378784,
+ "eval_mean_token_accuracy": 0.8282468245270547,
+ "eval_num_tokens": 7415795.0,
+ "eval_runtime": 98.5199,
+ "eval_samples_per_second": 14.413,
+ "eval_steps_per_second": 1.807,
+ "step": 3040
+ },
+ {
+ "entropy": 0.2161025382578373,
+ "epoch": 7.37394451145959,
+ "grad_norm": 0.802244246006012,
+ "learning_rate": 1.418071875579028e-05,
+ "loss": 0.14670108556747435,
+ "mean_token_accuracy": 0.9525101512670517,
+ "num_tokens": 7464083.0,
+ "step": 3060
+ },
+ {
+ "epoch": 7.37394451145959,
+ "eval_entropy": 0.3249504163526417,
+ "eval_loss": 0.8168519139289856,
+ "eval_mean_token_accuracy": 0.8281265336476015,
+ "eval_num_tokens": 7464083.0,
+ "eval_runtime": 98.4837,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 3060
+ },
+ {
+ "entropy": 0.21994274333119393,
+ "epoch": 7.422195416164053,
+ "grad_norm": 0.847614586353302,
+ "learning_rate": 1.370094111838762e-05,
+ "loss": 0.14627058506011964,
+ "mean_token_accuracy": 0.9523571625351905,
+ "num_tokens": 7513568.0,
+ "step": 3080
+ },
+ {
+ "epoch": 7.422195416164053,
+ "eval_entropy": 0.3224444846423824,
+ "eval_loss": 0.8206745982170105,
+ "eval_mean_token_accuracy": 0.8285097687431936,
+ "eval_num_tokens": 7513568.0,
+ "eval_runtime": 98.5699,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 3080
+ },
+ {
+ "entropy": 0.21474989503622055,
+ "epoch": 7.470446320868517,
+ "grad_norm": 0.8615112900733948,
+ "learning_rate": 1.3227515017856523e-05,
+ "loss": 0.14166855812072754,
+ "mean_token_accuracy": 0.9545643255114555,
+ "num_tokens": 7560503.0,
+ "step": 3100
+ },
+ {
+ "epoch": 7.470446320868517,
+ "eval_entropy": 0.3131480763634939,
+ "eval_loss": 0.8318397402763367,
+ "eval_mean_token_accuracy": 0.8293128348468395,
+ "eval_num_tokens": 7560503.0,
+ "eval_runtime": 98.5748,
+ "eval_samples_per_second": 14.405,
+ "eval_steps_per_second": 1.806,
+ "step": 3100
+ },
+ {
+ "entropy": 0.20732707269489764,
+ "epoch": 7.518697225572979,
+ "grad_norm": 0.8595603108406067,
+ "learning_rate": 1.2760574428183327e-05,
+ "loss": 0.1390355110168457,
+ "mean_token_accuracy": 0.9545612305402755,
+ "num_tokens": 7610472.0,
+ "step": 3120
+ },
+ {
+ "epoch": 7.518697225572979,
+ "eval_entropy": 0.32303216425555475,
+ "eval_loss": 0.8206163048744202,
+ "eval_mean_token_accuracy": 0.8276272057817223,
+ "eval_num_tokens": 7610472.0,
+ "eval_runtime": 98.6296,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 3120
+ },
+ {
+ "entropy": 0.20272110737860202,
+ "epoch": 7.566948130277443,
+ "grad_norm": 0.9171860814094543,
+ "learning_rate": 1.2300251488031559e-05,
+ "loss": 0.13609397411346436,
+ "mean_token_accuracy": 0.9557161509990693,
+ "num_tokens": 7662595.0,
+ "step": 3140
+ },
+ {
+ "epoch": 7.566948130277443,
+ "eval_entropy": 0.31589331360680334,
+ "eval_loss": 0.8248094916343689,
+ "eval_mean_token_accuracy": 0.8292221159077762,
+ "eval_num_tokens": 7662595.0,
+ "eval_runtime": 98.4574,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 3140
+ },
+ {
+ "entropy": 0.21678446158766745,
+ "epoch": 7.615199034981906,
+ "grad_norm": 0.870278000831604,
+ "learning_rate": 1.1846676463348161e-05,
+ "loss": 0.14677901268005372,
+ "mean_token_accuracy": 0.9534812569618225,
+ "num_tokens": 7711331.0,
+ "step": 3160
+ },
+ {
+ "epoch": 7.615199034981906,
+ "eval_entropy": 0.3183298571391052,
+ "eval_loss": 0.8235951662063599,
+ "eval_mean_token_accuracy": 0.8291523583819357,
+ "eval_num_tokens": 7711331.0,
+ "eval_runtime": 98.5743,
+ "eval_samples_per_second": 14.405,
+ "eval_steps_per_second": 1.806,
+ "step": 3160
+ },
+ {
+ "entropy": 0.2208289708942175,
+ "epoch": 7.66344993968637,
+ "grad_norm": 0.8425227999687195,
+ "learning_rate": 1.1399977710499864e-05,
+ "loss": 0.146539568901062,
+ "mean_token_accuracy": 0.9524248003959656,
+ "num_tokens": 7758759.0,
+ "step": 3180
+ },
+ {
+ "epoch": 7.66344993968637,
+ "eval_entropy": 0.3229933256346188,
+ "eval_loss": 0.8207331299781799,
+ "eval_mean_token_accuracy": 0.8284426217668512,
+ "eval_num_tokens": 7758759.0,
+ "eval_runtime": 98.6951,
+ "eval_samples_per_second": 14.388,
+ "eval_steps_per_second": 1.804,
+ "step": 3180
+ },
+ {
+ "entropy": 0.2222116433084011,
+ "epoch": 7.711700844390832,
+ "grad_norm": 0.8922525644302368,
+ "learning_rate": 1.0960281639949764e-05,
+ "loss": 0.14516948461532592,
+ "mean_token_accuracy": 0.952321757376194,
+ "num_tokens": 7805791.0,
+ "step": 3200
+ },
+ {
+ "epoch": 7.711700844390832,
+ "eval_entropy": 0.32119952912410993,
+ "eval_loss": 0.8213328123092651,
+ "eval_mean_token_accuracy": 0.8285770188556628,
+ "eval_num_tokens": 7805791.0,
+ "eval_runtime": 98.6437,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 3200
+ },
+ {
+ "entropy": 0.21560499407351016,
+ "epoch": 7.759951749095295,
+ "grad_norm": 0.9485028386116028,
+ "learning_rate": 1.0527712680484752e-05,
+ "loss": 0.1465426802635193,
+ "mean_token_accuracy": 0.9517758965492249,
+ "num_tokens": 7854924.0,
+ "step": 3220
+ },
+ {
+ "epoch": 7.759951749095295,
+ "eval_entropy": 0.3180267366298129,
+ "eval_loss": 0.8266745805740356,
+ "eval_mean_token_accuracy": 0.8283248446630628,
+ "eval_num_tokens": 7854924.0,
+ "eval_runtime": 98.5477,
+ "eval_samples_per_second": 14.409,
+ "eval_steps_per_second": 1.806,
+ "step": 3220
+ },
+ {
+ "entropy": 0.20934636779129506,
+ "epoch": 7.808202653799759,
+ "grad_norm": 0.7967023253440857,
+ "learning_rate": 1.0102393244003528e-05,
+ "loss": 0.14084668159484864,
+ "mean_token_accuracy": 0.9531109407544136,
+ "num_tokens": 7905721.0,
+ "step": 3240
+ },
+ {
+ "epoch": 7.808202653799759,
+ "eval_entropy": 0.31779277040047593,
+ "eval_loss": 0.8262729048728943,
+ "eval_mean_token_accuracy": 0.8288095492995187,
+ "eval_num_tokens": 7905721.0,
+ "eval_runtime": 98.6036,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 3240
+ },
+ {
+ "entropy": 0.215614727884531,
+ "epoch": 7.856453558504222,
+ "grad_norm": 0.7962915301322937,
+ "learning_rate": 9.684443690875602e-06,
+ "loss": 0.14136850833892822,
+ "mean_token_accuracy": 0.9531293407082557,
+ "num_tokens": 7955010.0,
+ "step": 3260
+ },
+ {
+ "epoch": 7.856453558504222,
+ "eval_entropy": 0.31656171481930806,
+ "eval_loss": 0.8303171992301941,
+ "eval_mean_token_accuracy": 0.8285069612974532,
+ "eval_num_tokens": 7955010.0,
+ "eval_runtime": 98.5105,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 3260
+ },
+ {
+ "entropy": 0.21908088140189647,
+ "epoch": 7.904704463208685,
+ "grad_norm": 0.9550889134407043,
+ "learning_rate": 9.273982295880618e-06,
+ "loss": 0.14732669591903685,
+ "mean_token_accuracy": 0.9522618219256401,
+ "num_tokens": 8001797.0,
+ "step": 3280
+ },
+ {
+ "epoch": 7.904704463208685,
+ "eval_entropy": 0.3193875088283185,
+ "eval_loss": 0.8240101933479309,
+ "eval_mean_token_accuracy": 0.8285468552219734,
+ "eval_num_tokens": 8001797.0,
+ "eval_runtime": 98.4506,
+ "eval_samples_per_second": 14.423,
+ "eval_steps_per_second": 1.808,
+ "step": 3280
+ },
+ {
+ "entropy": 0.21472855806350707,
+ "epoch": 7.952955367913148,
+ "grad_norm": 0.8032136559486389,
+ "learning_rate": 8.871125214738089e-06,
+ "loss": 0.1441211223602295,
+ "mean_token_accuracy": 0.9532426640391349,
+ "num_tokens": 8050462.0,
+ "step": 3300
+ },
+ {
+ "epoch": 7.952955367913148,
+ "eval_entropy": 0.3145652976766061,
+ "eval_loss": 0.8368120789527893,
+ "eval_mean_token_accuracy": 0.8285725685987579,
+ "eval_num_tokens": 8050462.0,
+ "eval_runtime": 98.5667,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 3300
+ },
+ {
+ "entropy": 0.20777193781657097,
+ "epoch": 8.0,
+ "grad_norm": 1.9440633058547974,
+ "learning_rate": 8.475986451236623e-06,
+ "loss": 0.13878930807113649,
+ "mean_token_accuracy": 0.9542667560088329,
+ "num_tokens": 8098232.0,
+ "step": 3320
+ },
+ {
+ "epoch": 8.0,
+ "eval_entropy": 0.3138233663493328,
+ "eval_loss": 0.8311719298362732,
+ "eval_mean_token_accuracy": 0.8295681362741449,
+ "eval_num_tokens": 8098232.0,
+ "eval_runtime": 98.401,
+ "eval_samples_per_second": 14.431,
+ "eval_steps_per_second": 1.809,
+ "step": 3320
+ },
+ {
+ "entropy": 0.19024005569517613,
+ "epoch": 8.048250904704464,
+ "grad_norm": 0.6860044598579407,
+ "learning_rate": 8.088677824972262e-06,
+ "loss": 0.104938805103302,
+ "mean_token_accuracy": 0.9669251158833504,
+ "num_tokens": 8145004.0,
+ "step": 3340
+ },
+ {
+ "epoch": 8.048250904704464,
+ "eval_entropy": 0.2925276593043563,
+ "eval_loss": 0.9170228838920593,
+ "eval_mean_token_accuracy": 0.8260884147681548,
+ "eval_num_tokens": 8145004.0,
+ "eval_runtime": 98.3738,
+ "eval_samples_per_second": 14.435,
+ "eval_steps_per_second": 1.809,
+ "step": 3340
+ },
+ {
+ "entropy": 0.17429662458598613,
+ "epoch": 8.096501809408926,
+ "grad_norm": 0.6853544116020203,
+ "learning_rate": 7.70930893970481e-06,
+ "loss": 0.09820220470428467,
+ "mean_token_accuracy": 0.9690332591533661,
+ "num_tokens": 8193186.0,
+ "step": 3360
+ },
+ {
+ "epoch": 8.096501809408926,
+ "eval_entropy": 0.29154844699280985,
+ "eval_loss": 0.9182955026626587,
+ "eval_mean_token_accuracy": 0.8260592245653774,
+ "eval_num_tokens": 8193186.0,
+ "eval_runtime": 98.5024,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 3360
+ },
+ {
+ "entropy": 0.17998851463198662,
+ "epoch": 8.14475271411339,
+ "grad_norm": 0.6730113625526428,
+ "learning_rate": 7.337987152341293e-06,
+ "loss": 0.10509306192398071,
+ "mean_token_accuracy": 0.9663080736994744,
+ "num_tokens": 8241385.0,
+ "step": 3380
+ },
+ {
+ "epoch": 8.14475271411339,
+ "eval_entropy": 0.29224191346530165,
+ "eval_loss": 0.9142967462539673,
+ "eval_mean_token_accuracy": 0.825743991672323,
+ "eval_num_tokens": 8241385.0,
+ "eval_runtime": 98.3875,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 3380
+ },
+ {
+ "entropy": 0.16425799671560526,
+ "epoch": 8.193003618817853,
+ "grad_norm": 0.8334925174713135,
+ "learning_rate": 6.974817542555239e-06,
+ "loss": 0.09438974857330322,
+ "mean_token_accuracy": 0.9704208940267562,
+ "num_tokens": 8294458.0,
+ "step": 3400
+ },
+ {
+ "epoch": 8.193003618817853,
+ "eval_entropy": 0.2905296404542548,
+ "eval_loss": 0.9226241707801819,
+ "eval_mean_token_accuracy": 0.8262435773785195,
+ "eval_num_tokens": 8294458.0,
+ "eval_runtime": 98.6148,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 3400
+ },
+ {
+ "entropy": 0.16865960750728845,
+ "epoch": 8.241254523522317,
+ "grad_norm": 0.7445836663246155,
+ "learning_rate": 6.619902883050312e-06,
+ "loss": 0.09588083028793334,
+ "mean_token_accuracy": 0.9690534204244614,
+ "num_tokens": 8347663.0,
+ "step": 3420
+ },
+ {
+ "epoch": 8.241254523522317,
+ "eval_entropy": 0.28748046489578954,
+ "eval_loss": 0.9263729453086853,
+ "eval_mean_token_accuracy": 0.8265480908115258,
+ "eval_num_tokens": 8347663.0,
+ "eval_runtime": 98.5098,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 3420
+ },
+ {
+ "entropy": 0.1850855764001608,
+ "epoch": 8.289505428226779,
+ "grad_norm": 0.8839195370674133,
+ "learning_rate": 6.273343610476938e-06,
+ "loss": 0.1067918062210083,
+ "mean_token_accuracy": 0.9668746247887612,
+ "num_tokens": 8391318.0,
+ "step": 3440
+ },
+ {
+ "epoch": 8.289505428226779,
+ "eval_entropy": 0.2879818446850509,
+ "eval_loss": 0.9265933036804199,
+ "eval_mean_token_accuracy": 0.826676292700714,
+ "eval_num_tokens": 8391318.0,
+ "eval_runtime": 98.491,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 3440
+ },
+ {
+ "entropy": 0.1764491790905595,
+ "epoch": 8.337756332931242,
+ "grad_norm": 0.8340020775794983,
+ "learning_rate": 5.935237797009842e-06,
+ "loss": 0.10066965818405152,
+ "mean_token_accuracy": 0.9678841292858124,
+ "num_tokens": 8440740.0,
+ "step": 3460
+ },
+ {
+ "epoch": 8.337756332931242,
+ "eval_entropy": 0.28914196303721224,
+ "eval_loss": 0.9251464605331421,
+ "eval_mean_token_accuracy": 0.8263801664448855,
+ "eval_num_tokens": 8440740.0,
+ "eval_runtime": 98.366,
+ "eval_samples_per_second": 14.436,
+ "eval_steps_per_second": 1.81,
+ "step": 3460
+ },
+ {
+ "entropy": 0.17226074449717999,
+ "epoch": 8.386007237635706,
+ "grad_norm": 0.7149388194084167,
+ "learning_rate": 5.605681122594823e-06,
+ "loss": 0.10052882432937622,
+ "mean_token_accuracy": 0.9679374754428863,
+ "num_tokens": 8490767.0,
+ "step": 3480
+ },
+ {
+ "epoch": 8.386007237635706,
+ "eval_entropy": 0.28626472519689733,
+ "eval_loss": 0.9307431578636169,
+ "eval_mean_token_accuracy": 0.8269837619213576,
+ "eval_num_tokens": 8490767.0,
+ "eval_runtime": 98.3539,
+ "eval_samples_per_second": 14.438,
+ "eval_steps_per_second": 1.81,
+ "step": 3480
+ },
+ {
+ "entropy": 0.17341369036585091,
+ "epoch": 8.43425814234017,
+ "grad_norm": 0.6688793301582336,
+ "learning_rate": 5.28476684787248e-06,
+ "loss": 0.10054340362548828,
+ "mean_token_accuracy": 0.9675475805997849,
+ "num_tokens": 8538959.0,
+ "step": 3500
+ },
+ {
+ "epoch": 8.43425814234017,
+ "eval_entropy": 0.28786867212378575,
+ "eval_loss": 0.9303860664367676,
+ "eval_mean_token_accuracy": 0.8259982466697693,
+ "eval_num_tokens": 8538959.0,
+ "eval_runtime": 98.627,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 3500
+ },
+ {
+ "entropy": 0.16994166951626538,
+ "epoch": 8.482509047044632,
+ "grad_norm": 0.7877247333526611,
+ "learning_rate": 4.972585787786484e-06,
+ "loss": 0.0993818998336792,
+ "mean_token_accuracy": 0.9688723921775818,
+ "num_tokens": 8587583.0,
+ "step": 3520
+ },
+ {
+ "epoch": 8.482509047044632,
+ "eval_entropy": 0.287084077600013,
+ "eval_loss": 0.9293763041496277,
+ "eval_mean_token_accuracy": 0.8268557224380836,
+ "eval_num_tokens": 8587583.0,
+ "eval_runtime": 98.5488,
+ "eval_samples_per_second": 14.409,
+ "eval_steps_per_second": 1.806,
+ "step": 3520
+ },
+ {
+ "entropy": 0.1698782866820693,
+ "epoch": 8.530759951749095,
+ "grad_norm": 0.6359167695045471,
+ "learning_rate": 4.66922628588412e-06,
+ "loss": 0.10011144876480102,
+ "mean_token_accuracy": 0.968739403784275,
+ "num_tokens": 8636960.0,
+ "step": 3540
+ },
+ {
+ "epoch": 8.530759951749095,
+ "eval_entropy": 0.2874377721481109,
+ "eval_loss": 0.9322494268417358,
+ "eval_mean_token_accuracy": 0.8263491124249576,
+ "eval_num_tokens": 8636960.0,
+ "eval_runtime": 98.3916,
+ "eval_samples_per_second": 14.432,
+ "eval_steps_per_second": 1.809,
+ "step": 3540
+ },
+ {
+ "entropy": 0.17174113132059574,
+ "epoch": 8.579010856453559,
+ "grad_norm": 0.7202840447425842,
+ "learning_rate": 4.374774189316017e-06,
+ "loss": 0.10309455394744874,
+ "mean_token_accuracy": 0.9690345630049706,
+ "num_tokens": 8685482.0,
+ "step": 3560
+ },
+ {
+ "epoch": 8.579010856453559,
+ "eval_entropy": 0.286078037673168,
+ "eval_loss": 0.9365394115447998,
+ "eval_mean_token_accuracy": 0.8265787784972888,
+ "eval_num_tokens": 8685482.0,
+ "eval_runtime": 98.5151,
+ "eval_samples_per_second": 14.414,
+ "eval_steps_per_second": 1.807,
+ "step": 3560
+ },
+ {
+ "entropy": 0.1779118437319994,
+ "epoch": 8.627261761158021,
+ "grad_norm": 0.7755381464958191,
+ "learning_rate": 4.08931282454246e-06,
+ "loss": 0.10361287593841553,
+ "mean_token_accuracy": 0.9668692052364349,
+ "num_tokens": 8733512.0,
+ "step": 3580
+ },
+ {
+ "epoch": 8.627261761158021,
+ "eval_entropy": 0.2901240846749102,
+ "eval_loss": 0.9282606244087219,
+ "eval_mean_token_accuracy": 0.8257483565405513,
+ "eval_num_tokens": 8733512.0,
+ "eval_runtime": 98.3953,
+ "eval_samples_per_second": 14.432,
+ "eval_steps_per_second": 1.809,
+ "step": 3580
+ },
+ {
+ "entropy": 0.17399974688887596,
+ "epoch": 8.675512665862485,
+ "grad_norm": 0.6068329215049744,
+ "learning_rate": 3.8129229737530146e-06,
+ "loss": 0.10165332555770874,
+ "mean_token_accuracy": 0.9683542385697365,
+ "num_tokens": 8781726.0,
+ "step": 3600
+ },
+ {
+ "epoch": 8.675512665862485,
+ "eval_entropy": 0.28656756911384923,
+ "eval_loss": 0.93599933385849,
+ "eval_mean_token_accuracy": 0.8266928818788422,
+ "eval_num_tokens": 8781726.0,
+ "eval_runtime": 98.4991,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 3600
+ },
+ {
+ "entropy": 0.17109180577099323,
+ "epoch": 8.723763570566948,
+ "grad_norm": 0.671509325504303,
+ "learning_rate": 3.5456828520060676e-06,
+ "loss": 0.10388398170471191,
+ "mean_token_accuracy": 0.9666652023792267,
+ "num_tokens": 8831964.0,
+ "step": 3620
+ },
+ {
+ "epoch": 8.723763570566948,
+ "eval_entropy": 0.28785753225007754,
+ "eval_loss": 0.9345396161079407,
+ "eval_mean_token_accuracy": 0.8258346751164855,
+ "eval_num_tokens": 8831964.0,
+ "eval_runtime": 98.5634,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 3620
+ },
+ {
+ "entropy": 0.17080388870090246,
+ "epoch": 8.772014475271412,
+ "grad_norm": 0.9064943194389343,
+ "learning_rate": 3.2876680850949342e-06,
+ "loss": 0.09955486059188842,
+ "mean_token_accuracy": 0.9688052341341973,
+ "num_tokens": 8881274.0,
+ "step": 3640
+ },
+ {
+ "epoch": 8.772014475271412,
+ "eval_entropy": 0.2863340745313784,
+ "eval_loss": 0.9368643760681152,
+ "eval_mean_token_accuracy": 0.826385034604019,
+ "eval_num_tokens": 8881274.0,
+ "eval_runtime": 98.4956,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 3640
+ },
+ {
+ "entropy": 0.1739195568487048,
+ "epoch": 8.820265379975874,
+ "grad_norm": 0.645410418510437,
+ "learning_rate": 3.0389516881466266e-06,
+ "loss": 0.1038020372390747,
+ "mean_token_accuracy": 0.965588228404522,
+ "num_tokens": 8928303.0,
+ "step": 3660
+ },
+ {
+ "epoch": 8.820265379975874,
+ "eval_entropy": 0.2869644621282481,
+ "eval_loss": 0.9337670803070068,
+ "eval_mean_token_accuracy": 0.8266236021947325,
+ "eval_num_tokens": 8928303.0,
+ "eval_runtime": 98.3869,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 3660
+ },
+ {
+ "entropy": 0.16863423213362694,
+ "epoch": 8.868516284680338,
+ "grad_norm": 0.6110771298408508,
+ "learning_rate": 2.7996040449594445e-06,
+ "loss": 0.10157880783081055,
+ "mean_token_accuracy": 0.967547157406807,
+ "num_tokens": 8977325.0,
+ "step": 3680
+ },
+ {
+ "epoch": 8.868516284680338,
+ "eval_entropy": 0.28859044577968257,
+ "eval_loss": 0.9297890067100525,
+ "eval_mean_token_accuracy": 0.8264903181724335,
+ "eval_num_tokens": 8977325.0,
+ "eval_runtime": 98.5054,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 3680
+ },
+ {
+ "entropy": 0.1767999932169914,
+ "epoch": 8.916767189384801,
+ "grad_norm": 0.6543433666229248,
+ "learning_rate": 2.5696928880852117e-06,
+ "loss": 0.10324364900588989,
+ "mean_token_accuracy": 0.9673267215490341,
+ "num_tokens": 9026898.0,
+ "step": 3700
+ },
+ {
+ "epoch": 8.916767189384801,
+ "eval_entropy": 0.2885892426364877,
+ "eval_loss": 0.928764820098877,
+ "eval_mean_token_accuracy": 0.8265983676642514,
+ "eval_num_tokens": 9026898.0,
+ "eval_runtime": 98.4914,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 3700
+ },
+ {
+ "entropy": 0.16943628452718257,
+ "epoch": 8.965018094089265,
+ "grad_norm": 0.8403868079185486,
+ "learning_rate": 2.3492832796617217e-06,
+ "loss": 0.09948049783706665,
+ "mean_token_accuracy": 0.967706835269928,
+ "num_tokens": 9077388.0,
+ "step": 3720
+ },
+ {
+ "epoch": 8.965018094089265,
+ "eval_entropy": 0.2884856275795551,
+ "eval_loss": 0.933276355266571,
+ "eval_mean_token_accuracy": 0.8261616119507993,
+ "eval_num_tokens": 9077388.0,
+ "eval_runtime": 98.6259,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 3720
+ },
+ {
+ "entropy": 0.17359357613783616,
+ "epoch": 9.012062726176115,
+ "grad_norm": 0.37213996052742004,
+ "learning_rate": 2.138437593000977e-06,
+ "loss": 0.10089871883392335,
+ "mean_token_accuracy": 0.9696638874518566,
+ "num_tokens": 9122597.0,
+ "step": 3740
+ },
+ {
+ "epoch": 9.012062726176115,
+ "eval_entropy": 0.2875421148505104,
+ "eval_loss": 0.9357590079307556,
+ "eval_mean_token_accuracy": 0.8264994859025719,
+ "eval_num_tokens": 9122597.0,
+ "eval_runtime": 98.523,
+ "eval_samples_per_second": 14.413,
+ "eval_steps_per_second": 1.807,
+ "step": 3740
+ },
+ {
+ "entropy": 0.1642639346420765,
+ "epoch": 9.060313630880579,
+ "grad_norm": 0.5511618852615356,
+ "learning_rate": 1.937215494938237e-06,
+ "loss": 0.08524587750434875,
+ "mean_token_accuracy": 0.9754035726189614,
+ "num_tokens": 9171477.0,
+ "step": 3760
+ },
+ {
+ "epoch": 9.060313630880579,
+ "eval_entropy": 0.2820046850469675,
+ "eval_loss": 0.9588018655776978,
+ "eval_mean_token_accuracy": 0.8257443647036392,
+ "eval_num_tokens": 9171477.0,
+ "eval_runtime": 98.3992,
+ "eval_samples_per_second": 14.431,
+ "eval_steps_per_second": 1.809,
+ "step": 3760
+ },
+ {
+ "entropy": 0.1630103051662445,
+ "epoch": 9.108564535585042,
+ "grad_norm": 0.5125291347503662,
+ "learning_rate": 1.7456739289470143e-06,
+ "loss": 0.0895907461643219,
+ "mean_token_accuracy": 0.973100857436657,
+ "num_tokens": 9217543.0,
+ "step": 3780
+ },
+ {
+ "epoch": 9.108564535585042,
+ "eval_entropy": 0.2786683516388529,
+ "eval_loss": 0.9743365049362183,
+ "eval_mean_token_accuracy": 0.8253507131940863,
+ "eval_num_tokens": 9217543.0,
+ "eval_runtime": 98.3882,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 3780
+ },
+ {
+ "entropy": 0.15873534101992845,
+ "epoch": 9.156815440289506,
+ "grad_norm": 0.5697775483131409,
+ "learning_rate": 1.5638670990247788e-06,
+ "loss": 0.08552997708320617,
+ "mean_token_accuracy": 0.973672679066658,
+ "num_tokens": 9268724.0,
+ "step": 3800
+ },
+ {
+ "epoch": 9.156815440289506,
+ "eval_entropy": 0.2781428360704626,
+ "eval_loss": 0.9790021777153015,
+ "eval_mean_token_accuracy": 0.8251898780967413,
+ "eval_num_tokens": 9268724.0,
+ "eval_runtime": 98.5449,
+ "eval_samples_per_second": 14.41,
+ "eval_steps_per_second": 1.806,
+ "step": 3800
+ },
+ {
+ "entropy": 0.1606229657307267,
+ "epoch": 9.205066344993968,
+ "grad_norm": 0.5715742707252502,
+ "learning_rate": 1.3918464543538225e-06,
+ "loss": 0.08984670042991638,
+ "mean_token_accuracy": 0.9729854613542557,
+ "num_tokens": 9316805.0,
+ "step": 3820
+ },
+ {
+ "epoch": 9.205066344993968,
+ "eval_entropy": 0.2777416408899125,
+ "eval_loss": 0.9808777570724487,
+ "eval_mean_token_accuracy": 0.8252255926641185,
+ "eval_num_tokens": 9316805.0,
+ "eval_runtime": 98.627,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 3820
+ },
+ {
+ "entropy": 0.15027269292622805,
+ "epoch": 9.253317249698432,
+ "grad_norm": 0.3359828293323517,
+ "learning_rate": 1.229660674741786e-06,
+ "loss": 0.08229721784591675,
+ "mean_token_accuracy": 0.9754723474383354,
+ "num_tokens": 9369940.0,
+ "step": 3840
+ },
+ {
+ "epoch": 9.253317249698432,
+ "eval_entropy": 0.2768587637483404,
+ "eval_loss": 0.9837455749511719,
+ "eval_mean_token_accuracy": 0.825493957889214,
+ "eval_num_tokens": 9369940.0,
+ "eval_runtime": 98.534,
+ "eval_samples_per_second": 14.411,
+ "eval_steps_per_second": 1.806,
+ "step": 3840
+ },
+ {
+ "entropy": 0.16767099779099226,
+ "epoch": 9.301568154402895,
+ "grad_norm": 0.5708875060081482,
+ "learning_rate": 1.077355656845822e-06,
+ "loss": 0.08888025283813476,
+ "mean_token_accuracy": 0.9715678468346596,
+ "num_tokens": 9416788.0,
+ "step": 3860
+ },
+ {
+ "epoch": 9.301568154402895,
+ "eval_entropy": 0.27612990283229377,
+ "eval_loss": 0.9873999357223511,
+ "eval_mean_token_accuracy": 0.8252144948150335,
+ "eval_num_tokens": 9416788.0,
+ "eval_runtime": 98.3889,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 3860
+ },
+ {
+ "entropy": 0.1549418892711401,
+ "epoch": 9.349819059107359,
+ "grad_norm": 0.7601456046104431,
+ "learning_rate": 9.349745011844029e-07,
+ "loss": 0.08411069512367249,
+ "mean_token_accuracy": 0.9741103485226631,
+ "num_tokens": 9467862.0,
+ "step": 3880
+ },
+ {
+ "epoch": 9.349819059107359,
+ "eval_entropy": 0.2757877716689967,
+ "eval_loss": 0.9892058968544006,
+ "eval_mean_token_accuracy": 0.8250612017143978,
+ "eval_num_tokens": 9467862.0,
+ "eval_runtime": 98.3869,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 3880
+ },
+ {
+ "entropy": 0.16111517809331416,
+ "epoch": 9.398069963811821,
+ "grad_norm": 0.6774276494979858,
+ "learning_rate": 8.025574999403695e-07,
+ "loss": 0.08765592575073242,
+ "mean_token_accuracy": 0.9734174907207489,
+ "num_tokens": 9514611.0,
+ "step": 3900
+ },
+ {
+ "epoch": 9.398069963811821,
+ "eval_entropy": 0.275664389049739,
+ "eval_loss": 0.989257276058197,
+ "eval_mean_token_accuracy": 0.8252896914321385,
+ "eval_num_tokens": 9514611.0,
+ "eval_runtime": 98.4649,
+ "eval_samples_per_second": 14.421,
+ "eval_steps_per_second": 1.808,
+ "step": 3900
+ },
+ {
+ "entropy": 0.15701593235135078,
+ "epoch": 9.446320868516285,
+ "grad_norm": 0.6130298376083374,
+ "learning_rate": 6.801421255587377e-07,
+ "loss": 0.0853091835975647,
+ "mean_token_accuracy": 0.9740088582038879,
+ "num_tokens": 9563437.0,
+ "step": 3920
+ },
+ {
+ "epoch": 9.446320868516285,
+ "eval_entropy": 0.27559053881114787,
+ "eval_loss": 0.9901270866394043,
+ "eval_mean_token_accuracy": 0.8251507653948966,
+ "eval_num_tokens": 9563437.0,
+ "eval_runtime": 98.5285,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 3920
+ },
+ {
+ "entropy": 0.15848954878747462,
+ "epoch": 9.494571773220748,
+ "grad_norm": 0.5928868651390076,
+ "learning_rate": 5.677630201424204e-07,
+ "loss": 0.08467000126838684,
+ "mean_token_accuracy": 0.9725634664297104,
+ "num_tokens": 9614034.0,
+ "step": 3940
+ },
+ {
+ "epoch": 9.494571773220748,
+ "eval_entropy": 0.2750965692521481,
+ "eval_loss": 0.9914345145225525,
+ "eval_mean_token_accuracy": 0.8252089954494091,
+ "eval_num_tokens": 9614034.0,
+ "eval_runtime": 98.5049,
+ "eval_samples_per_second": 14.416,
+ "eval_steps_per_second": 1.807,
+ "step": 3940
+ },
+ {
+ "entropy": 0.15644967313855887,
+ "epoch": 9.54282267792521,
+ "grad_norm": 0.5718610882759094,
+ "learning_rate": 4.6545198564893463e-07,
+ "loss": 0.08564762473106384,
+ "mean_token_accuracy": 0.9730244860053062,
+ "num_tokens": 9661266.0,
+ "step": 3960
+ },
+ {
+ "epoch": 9.54282267792521,
+ "eval_entropy": 0.27522430820076654,
+ "eval_loss": 0.9908939003944397,
+ "eval_mean_token_accuracy": 0.8251052994406625,
+ "eval_num_tokens": 9661266.0,
+ "eval_runtime": 98.3637,
+ "eval_samples_per_second": 14.436,
+ "eval_steps_per_second": 1.81,
+ "step": 3960
+ },
+ {
+ "entropy": 0.15196608882397414,
+ "epoch": 9.591073582629674,
+ "grad_norm": 0.752661406993866,
+ "learning_rate": 3.732379748908161e-07,
+ "loss": 0.08072127103805542,
+ "mean_token_accuracy": 0.9749975100159645,
+ "num_tokens": 9710458.0,
+ "step": 3980
+ },
+ {
+ "epoch": 9.591073582629674,
+ "eval_entropy": 0.2750371985890892,
+ "eval_loss": 0.9913897514343262,
+ "eval_mean_token_accuracy": 0.8248539731073915,
+ "eval_num_tokens": 9710458.0,
+ "eval_runtime": 98.4959,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 3980
+ },
+ {
+ "entropy": 0.16616767533123494,
+ "epoch": 9.639324487334138,
+ "grad_norm": 0.7107775211334229,
+ "learning_rate": 2.911470833423274e-07,
+ "loss": 0.08988997340202332,
+ "mean_token_accuracy": 0.9714749112725258,
+ "num_tokens": 9755797.0,
+ "step": 4000
+ },
+ {
+ "epoch": 9.639324487334138,
+ "eval_entropy": 0.27523158457172053,
+ "eval_loss": 0.9913759827613831,
+ "eval_mean_token_accuracy": 0.8250464505693885,
+ "eval_num_tokens": 9755797.0,
+ "eval_runtime": 98.5067,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 4000
+ },
+ {
+ "entropy": 0.15760798901319503,
+ "epoch": 9.687575392038601,
+ "grad_norm": 0.5563387870788574,
+ "learning_rate": 2.1920254175472834e-07,
+ "loss": 0.0877057135105133,
+ "mean_token_accuracy": 0.9734948605298996,
+ "num_tokens": 9805552.0,
+ "step": 4020
+ },
+ {
+ "epoch": 9.687575392038601,
+ "eval_entropy": 0.27528667726208655,
+ "eval_loss": 0.9914706349372864,
+ "eval_mean_token_accuracy": 0.8250808802883277,
+ "eval_num_tokens": 9805552.0,
+ "eval_runtime": 98.475,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 4020
+ },
+ {
+ "entropy": 0.161653464846313,
+ "epoch": 9.735826296743063,
+ "grad_norm": 0.5896415114402771,
+ "learning_rate": 1.5742470958227596e-07,
+ "loss": 0.08775997757911683,
+ "mean_token_accuracy": 0.9718310445547104,
+ "num_tokens": 9852843.0,
+ "step": 4040
+ },
+ {
+ "epoch": 9.735826296743063,
+ "eval_entropy": 0.27531274182073184,
+ "eval_loss": 0.9915319085121155,
+ "eval_mean_token_accuracy": 0.825061980258213,
+ "eval_num_tokens": 9852843.0,
+ "eval_runtime": 98.3873,
+ "eval_samples_per_second": 14.433,
+ "eval_steps_per_second": 1.809,
+ "step": 4040
+ },
+ {
+ "entropy": 0.16010624952614308,
+ "epoch": 9.784077201447527,
+ "grad_norm": 0.6022347211837769,
+ "learning_rate": 1.058310692207063e-07,
+ "loss": 0.08761411905288696,
+ "mean_token_accuracy": 0.9729624018073082,
+ "num_tokens": 9900151.0,
+ "step": 4060
+ },
+ {
+ "epoch": 9.784077201447527,
+ "eval_entropy": 0.27524732446737504,
+ "eval_loss": 0.991264820098877,
+ "eval_mean_token_accuracy": 0.8251047007153544,
+ "eval_num_tokens": 9900151.0,
+ "eval_runtime": 98.4834,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 4060
+ },
+ {
+ "entropy": 0.159021489135921,
+ "epoch": 9.83232810615199,
+ "grad_norm": 0.48451653122901917,
+ "learning_rate": 6.443622105994901e-08,
+ "loss": 0.08291574716567993,
+ "mean_token_accuracy": 0.9726693674921989,
+ "num_tokens": 9952926.0,
+ "step": 4080
+ },
+ {
+ "epoch": 9.83232810615199,
+ "eval_entropy": 0.27521907144717955,
+ "eval_loss": 0.9918096661567688,
+ "eval_mean_token_accuracy": 0.8250070894032382,
+ "eval_num_tokens": 9952926.0,
+ "eval_runtime": 98.3568,
+ "eval_samples_per_second": 14.437,
+ "eval_steps_per_second": 1.81,
+ "step": 4080
+ },
+ {
+ "entropy": 0.15175675600767136,
+ "epoch": 9.880579010856454,
+ "grad_norm": 0.5009680390357971,
+ "learning_rate": 3.325187935238231e-08,
+ "loss": 0.08214685916900635,
+ "mean_token_accuracy": 0.9752551406621933,
+ "num_tokens": 10003720.0,
+ "step": 4100
+ },
+ {
+ "epoch": 9.880579010856454,
+ "eval_entropy": 0.27522189488236826,
+ "eval_loss": 0.9918575882911682,
+ "eval_mean_token_accuracy": 0.8249149091458052,
+ "eval_num_tokens": 10003720.0,
+ "eval_runtime": 98.4802,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 4100
+ },
+ {
+ "entropy": 0.15018076542764902,
+ "epoch": 9.928829915560916,
+ "grad_norm": 0.6191889047622681,
+ "learning_rate": 1.2286868897828347e-08,
+ "loss": 0.08095031380653381,
+ "mean_token_accuracy": 0.9739995509386062,
+ "num_tokens": 10057590.0,
+ "step": 4120
+ },
+ {
+ "epoch": 9.928829915560916,
+ "eval_entropy": 0.27517637836464337,
+ "eval_loss": 0.991787850856781,
+ "eval_mean_token_accuracy": 0.8249709994605418,
+ "eval_num_tokens": 10057590.0,
+ "eval_runtime": 98.5572,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 4120
+ },
+ {
+ "entropy": 0.16240970119833947,
+ "epoch": 9.97708082026538,
+ "grad_norm": 0.5195524096488953,
+ "learning_rate": 1.5471225462640753e-09,
+ "loss": 0.08831822276115417,
+ "mean_token_accuracy": 0.9716430872678756,
+ "num_tokens": 10103443.0,
+ "step": 4140
+ },
+ {
+ "epoch": 9.97708082026538,
+ "eval_entropy": 0.27521504202250685,
+ "eval_loss": 0.9918516278266907,
+ "eval_mean_token_accuracy": 0.8251796190658313,
+ "eval_num_tokens": 10103443.0,
+ "eval_runtime": 98.3776,
+ "eval_samples_per_second": 14.434,
+ "eval_steps_per_second": 1.809,
+ "step": 4140
+ },
+ {
+ "epoch": 10.0,
+ "eval_entropy": 0.2751017221239176,
+ "eval_loss": 0.9913693070411682,
+ "eval_mean_token_accuracy": 0.8250626305515847,
+ "eval_num_tokens": 10122790.0,
+ "eval_runtime": 98.396,
+ "eval_samples_per_second": 14.431,
+ "eval_steps_per_second": 1.809,
+ "step": 4150
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": true
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 2.098801344154368e+18,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-420/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-420/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-420/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-420/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-420/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-420/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-420/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-420/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-420/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-420/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-420/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-420/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-420/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-420/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..e1ae955cd64f3af8d6bad3b8c9205f27937a24ac
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-420/trainer_state.json
@@ -0,0 +1,475 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.0120627261761157,
+ "eval_steps": 20,
+ "global_step": 420,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 2.1381586730112e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-440/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-440/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-440/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-440/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-440/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-440/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-440/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-440/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-440/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-440/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-440/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-440/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-440/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-440/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..3fff5f435b6cab928f5366ab412d941a8a454fe4
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-440/trainer_state.json
@@ -0,0 +1,496 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.060313630880579,
+ "eval_steps": 20,
+ "global_step": 440,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 2.25473118866688e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-460/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-460/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-460/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-460/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-460/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-460/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-460/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-460/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-460/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-460/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-460/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-460/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-460/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-460/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..2b8fe15c7b25a66d14c769eac692d1fb7097d1c2
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-460/trainer_state.json
@@ -0,0 +1,517 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.1085645355850422,
+ "eval_steps": 20,
+ "global_step": 460,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 2.34788114969856e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-480/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-480/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-480/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-480/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-480/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-480/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-480/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-480/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-480/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-480/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-480/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-480/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-480/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-480/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..19c3684404b9c98d9405093ac85bd40c8630ab0c
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-480/trainer_state.json
@@ -0,0 +1,538 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.1568154402895054,
+ "eval_steps": 20,
+ "global_step": 480,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 2.45117290591488e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-500/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-500/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-500/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-500/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-500/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-500/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-500/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-500/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-500/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-500/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-500/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-500/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-500/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-500/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..d61667eb6e42909d9d18cc8eff252abe668c902b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-500/trainer_state.json
@@ -0,0 +1,559 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.2050663449939687,
+ "eval_steps": 20,
+ "global_step": 500,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 2.54922621633792e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-520/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-520/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-520/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-520/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-520/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-520/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-520/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-520/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-520/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-520/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-520/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-520/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-520/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-520/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..511a49bd26c0e9ab580ce7fa8d82084196b24fce
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-520/trainer_state.json
@@ -0,0 +1,580 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.2533172496984317,
+ "eval_steps": 20,
+ "global_step": 520,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 2.65577319474432e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-540/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-540/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-540/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-540/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-540/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-540/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-540/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-540/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-540/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-540/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-540/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-540/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-540/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-540/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..9d6d91178d33e765c84fabec86851739ae97d183
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-540/trainer_state.json
@@ -0,0 +1,601 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.301568154402895,
+ "eval_steps": 20,
+ "global_step": 540,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 2.75268444192e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-560/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-560/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-560/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-560/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-560/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-560/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-560/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-560/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-560/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-560/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-560/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-560/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-560/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-560/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..81bf4b6d921d4a7beb27bf54ab3ef1f9b289a9be
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-560/trainer_state.json
@@ -0,0 +1,622 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.3498190591073582,
+ "eval_steps": 20,
+ "global_step": 560,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 2.84735943351552e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-580/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-580/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-580/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-580/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-580/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-580/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-580/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-580/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-580/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-580/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-580/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-580/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-580/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-580/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..7c7cc052c7fa4e6570df99b2e5adecc48dec4d97
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-580/trainer_state.json
@@ -0,0 +1,643 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.3980699638118215,
+ "eval_steps": 20,
+ "global_step": 580,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 2.93778075241728e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-60/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-60/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-60/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-60/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-60/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-60/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-60/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-60/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-60/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-60/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-60/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-60/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-60/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-60/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..763f0cca0e5a16acb1e02ce9e636a32bbb6a4b9d
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-60/trainer_state.json
@@ -0,0 +1,97 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.14475271411338964,
+ "eval_steps": 20,
+ "global_step": 60,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 3.0126534273024e+16,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-600/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-600/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-600/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-600/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-600/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-600/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-600/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-600/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-600/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-600/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-600/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-600/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-600/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-600/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..fd885e888f061e1b1100345f240cc7ee47667dc3
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-600/trainer_state.json
@@ -0,0 +1,664 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.4463208685162847,
+ "eval_steps": 20,
+ "global_step": 600,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 3.04147599205632e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-620/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-620/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-620/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-620/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-620/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-620/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-620/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-620/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-620/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-620/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-620/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-620/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-620/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-620/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..8e85926192e88e1f84c0277cfb30f51de078db95
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-620/trainer_state.json
@@ -0,0 +1,685 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.494571773220748,
+ "eval_steps": 20,
+ "global_step": 620,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 3.1453627153536e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-640/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-640/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-640/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-640/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-640/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-640/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-640/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-640/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-640/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-640/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-640/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-640/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-640/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-640/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..e351c7e2c382f30337784e78e3c50caaa8576b19
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-640/trainer_state.json
@@ -0,0 +1,706 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.5428226779252112,
+ "eval_steps": 20,
+ "global_step": 640,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 3.25206698390784e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-660/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-660/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-660/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-660/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-660/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-660/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-660/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-660/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-660/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-660/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-660/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-660/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-660/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-660/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..920e60c36ec33d0fc22d2c96b9b536911a89f5c1
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-660/trainer_state.json
@@ -0,0 +1,727 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.5910735826296745,
+ "eval_steps": 20,
+ "global_step": 660,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 3.34308326989056e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-680/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-680/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-680/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-680/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-680/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-680/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-680/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-680/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-680/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-680/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-680/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-680/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-680/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-680/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..2ae31455df44f729bb9d7cfa1c67786e285c5fd2
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-680/trainer_state.json
@@ -0,0 +1,748 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.6393244873341375,
+ "eval_steps": 20,
+ "global_step": 680,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 3.443393352e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-700/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-700/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-700/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-700/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-700/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-700/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-700/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-700/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-700/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-700/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-700/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-700/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-700/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-700/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..2a3c46ebd23b5109f41e44b6c4cd9c24181f86b2
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-700/trainer_state.json
@@ -0,0 +1,769 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.6875753920386007,
+ "eval_steps": 20,
+ "global_step": 700,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 3.5502822655104e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-720/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-720/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-720/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-720/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-720/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-720/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-720/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-720/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-720/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-720/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-720/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-720/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-720/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-720/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..3276455443ecb5882b880c253055de65b1385ace
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-720/trainer_state.json
@@ -0,0 +1,790 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.7358262967430638,
+ "eval_steps": 20,
+ "global_step": 720,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 3.66350381714688e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-740/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-740/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-740/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-740/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-740/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-740/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-740/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-740/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-740/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-740/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-740/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-740/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-740/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-740/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..1f61f827df821150ec4459de8186514c58a973e5
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-740/trainer_state.json
@@ -0,0 +1,811 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.784077201447527,
+ "eval_steps": 20,
+ "global_step": 740,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 3.76028512898304e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-760/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-760/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-760/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-760/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-760/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-760/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-760/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-760/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-760/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-760/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-760/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-760/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-760/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-760/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..3f686a0dc13c29d3e813a5034a20adad7cfa6f96
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-760/trainer_state.json
@@ -0,0 +1,832 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.8323281061519903,
+ "eval_steps": 20,
+ "global_step": 760,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ },
+ {
+ "entropy": 0.6141144648194313,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6467089056968689,
+ "learning_rate": 7.078840074092574e-05,
+ "loss": 0.5548617362976074,
+ "mean_token_accuracy": 0.8342376738786698,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.5951976082967908,
+ "eval_loss": 0.5850171446800232,
+ "eval_mean_token_accuracy": 0.8303041870004675,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 98.6601,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 760
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 3.8574083759232e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-780/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-780/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-780/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-780/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-780/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-780/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-780/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-780/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-780/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-780/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-780/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-780/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-780/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-780/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..41497cadd664d7842b5cfae211337401d3e4eaa4
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-780/trainer_state.json
@@ -0,0 +1,853 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.8805790108564535,
+ "eval_steps": 20,
+ "global_step": 780,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ },
+ {
+ "entropy": 0.6141144648194313,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6467089056968689,
+ "learning_rate": 7.078840074092574e-05,
+ "loss": 0.5548617362976074,
+ "mean_token_accuracy": 0.8342376738786698,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.5951976082967908,
+ "eval_loss": 0.5850171446800232,
+ "eval_mean_token_accuracy": 0.8303041870004675,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 98.6601,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 760
+ },
+ {
+ "entropy": 0.5961867064237595,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.41280242800712585,
+ "learning_rate": 7.061001336495465e-05,
+ "loss": 0.5511267185211182,
+ "mean_token_accuracy": 0.8392465204000473,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6167017998320333,
+ "eval_loss": 0.5826791524887085,
+ "eval_mean_token_accuracy": 0.8309678050239434,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 98.6564,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 780
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 3.96010516505856e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-80/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-80/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-80/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-80/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-80/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-80/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-80/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-80/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-80/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-80/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-80/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-80/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-80/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-80/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..2614808ce0d1cbbbf1ab81cf656f2a0d3d2508da
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-80/trainer_state.json
@@ -0,0 +1,118 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 0.19300361881785283,
+ "eval_steps": 20,
+ "global_step": 80,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 3.9943491108864e+16,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-800/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-800/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-800/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-800/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-800/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-800/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-800/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-800/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-800/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-800/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-800/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-800/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-800/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-800/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..d26c2f681e81deb982b2c246a034229827626f52
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-800/trainer_state.json
@@ -0,0 +1,874 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.9288299155609168,
+ "eval_steps": 20,
+ "global_step": 800,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ },
+ {
+ "entropy": 0.6141144648194313,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6467089056968689,
+ "learning_rate": 7.078840074092574e-05,
+ "loss": 0.5548617362976074,
+ "mean_token_accuracy": 0.8342376738786698,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.5951976082967908,
+ "eval_loss": 0.5850171446800232,
+ "eval_mean_token_accuracy": 0.8303041870004675,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 98.6601,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 760
+ },
+ {
+ "entropy": 0.5961867064237595,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.41280242800712585,
+ "learning_rate": 7.061001336495465e-05,
+ "loss": 0.5511267185211182,
+ "mean_token_accuracy": 0.8392465204000473,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6167017998320333,
+ "eval_loss": 0.5826791524887085,
+ "eval_mean_token_accuracy": 0.8309678050239434,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 98.6564,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 780
+ },
+ {
+ "entropy": 0.6025628983974457,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.4078947901725769,
+ "learning_rate": 7.042187293157881e-05,
+ "loss": 0.5552026271820069,
+ "mean_token_accuracy": 0.8359778299927711,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.6047008484601974,
+ "eval_loss": 0.5811123847961426,
+ "eval_mean_token_accuracy": 0.8320730804057603,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 98.6033,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 800
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 4.06481937130752e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-820/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-820/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-820/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-820/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-820/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-820/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-820/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-820/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-820/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-820/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-820/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-820/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-820/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-820/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..598cab60082bcc712dced435aef50ea8ca199d31
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-820/trainer_state.json
@@ -0,0 +1,895 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 1.97708082026538,
+ "eval_steps": 20,
+ "global_step": 820,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ },
+ {
+ "entropy": 0.6141144648194313,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6467089056968689,
+ "learning_rate": 7.078840074092574e-05,
+ "loss": 0.5548617362976074,
+ "mean_token_accuracy": 0.8342376738786698,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.5951976082967908,
+ "eval_loss": 0.5850171446800232,
+ "eval_mean_token_accuracy": 0.8303041870004675,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 98.6601,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 760
+ },
+ {
+ "entropy": 0.5961867064237595,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.41280242800712585,
+ "learning_rate": 7.061001336495465e-05,
+ "loss": 0.5511267185211182,
+ "mean_token_accuracy": 0.8392465204000473,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6167017998320333,
+ "eval_loss": 0.5826791524887085,
+ "eval_mean_token_accuracy": 0.8309678050239434,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 98.6564,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 780
+ },
+ {
+ "entropy": 0.6025628983974457,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.4078947901725769,
+ "learning_rate": 7.042187293157881e-05,
+ "loss": 0.5552026271820069,
+ "mean_token_accuracy": 0.8359778299927711,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.6047008484601974,
+ "eval_loss": 0.5811123847961426,
+ "eval_mean_token_accuracy": 0.8320730804057603,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 98.6033,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 800
+ },
+ {
+ "entropy": 0.6041995719075203,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.543890118598938,
+ "learning_rate": 7.022403268231535e-05,
+ "loss": 0.554384422302246,
+ "mean_token_accuracy": 0.8380729153752327,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.6005127917179901,
+ "eval_loss": 0.5800131559371948,
+ "eval_mean_token_accuracy": 0.8313186547059691,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 98.4119,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 820
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 4.15911823428864e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-840/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-840/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-840/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-840/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-840/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-840/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-840/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-840/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-840/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-840/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-840/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-840/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-840/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-840/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..582322764095d99326791f23e881bde1ca6daf2a
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-840/trainer_state.json
@@ -0,0 +1,916 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.0241254523522314,
+ "eval_steps": 20,
+ "global_step": 840,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ },
+ {
+ "entropy": 0.6141144648194313,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6467089056968689,
+ "learning_rate": 7.078840074092574e-05,
+ "loss": 0.5548617362976074,
+ "mean_token_accuracy": 0.8342376738786698,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.5951976082967908,
+ "eval_loss": 0.5850171446800232,
+ "eval_mean_token_accuracy": 0.8303041870004675,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 98.6601,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 760
+ },
+ {
+ "entropy": 0.5961867064237595,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.41280242800712585,
+ "learning_rate": 7.061001336495465e-05,
+ "loss": 0.5511267185211182,
+ "mean_token_accuracy": 0.8392465204000473,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6167017998320333,
+ "eval_loss": 0.5826791524887085,
+ "eval_mean_token_accuracy": 0.8309678050239434,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 98.6564,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 780
+ },
+ {
+ "entropy": 0.6025628983974457,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.4078947901725769,
+ "learning_rate": 7.042187293157881e-05,
+ "loss": 0.5552026271820069,
+ "mean_token_accuracy": 0.8359778299927711,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.6047008484601974,
+ "eval_loss": 0.5811123847961426,
+ "eval_mean_token_accuracy": 0.8320730804057603,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 98.6033,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 800
+ },
+ {
+ "entropy": 0.6041995719075203,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.543890118598938,
+ "learning_rate": 7.022403268231535e-05,
+ "loss": 0.554384422302246,
+ "mean_token_accuracy": 0.8380729153752327,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.6005127917179901,
+ "eval_loss": 0.5800131559371948,
+ "eval_mean_token_accuracy": 0.8313186547059691,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 98.4119,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 820
+ },
+ {
+ "entropy": 0.567442889396961,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.5015465021133423,
+ "learning_rate": 7.001654860361456e-05,
+ "loss": 0.5208251476287842,
+ "mean_token_accuracy": 0.8442784960453327,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.5785037809878253,
+ "eval_loss": 0.5865591168403625,
+ "eval_mean_token_accuracy": 0.8307993857378371,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 98.5061,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 840
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 4.26685771137024e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-860/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-860/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-860/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-860/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-860/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-860/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-860/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-860/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-860/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-860/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-860/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-860/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-860/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-860/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..319c439dfe0c70b5a345ca5c4123db515f15c45c
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-860/trainer_state.json
@@ -0,0 +1,937 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.0723763570566947,
+ "eval_steps": 20,
+ "global_step": 860,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ },
+ {
+ "entropy": 0.6141144648194313,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6467089056968689,
+ "learning_rate": 7.078840074092574e-05,
+ "loss": 0.5548617362976074,
+ "mean_token_accuracy": 0.8342376738786698,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.5951976082967908,
+ "eval_loss": 0.5850171446800232,
+ "eval_mean_token_accuracy": 0.8303041870004675,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 98.6601,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 760
+ },
+ {
+ "entropy": 0.5961867064237595,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.41280242800712585,
+ "learning_rate": 7.061001336495465e-05,
+ "loss": 0.5511267185211182,
+ "mean_token_accuracy": 0.8392465204000473,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6167017998320333,
+ "eval_loss": 0.5826791524887085,
+ "eval_mean_token_accuracy": 0.8309678050239434,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 98.6564,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 780
+ },
+ {
+ "entropy": 0.6025628983974457,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.4078947901725769,
+ "learning_rate": 7.042187293157881e-05,
+ "loss": 0.5552026271820069,
+ "mean_token_accuracy": 0.8359778299927711,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.6047008484601974,
+ "eval_loss": 0.5811123847961426,
+ "eval_mean_token_accuracy": 0.8320730804057603,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 98.6033,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 800
+ },
+ {
+ "entropy": 0.6041995719075203,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.543890118598938,
+ "learning_rate": 7.022403268231535e-05,
+ "loss": 0.554384422302246,
+ "mean_token_accuracy": 0.8380729153752327,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.6005127917179901,
+ "eval_loss": 0.5800131559371948,
+ "eval_mean_token_accuracy": 0.8313186547059691,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 98.4119,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 820
+ },
+ {
+ "entropy": 0.567442889396961,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.5015465021133423,
+ "learning_rate": 7.001654860361456e-05,
+ "loss": 0.5208251476287842,
+ "mean_token_accuracy": 0.8442784960453327,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.5785037809878253,
+ "eval_loss": 0.5865591168403625,
+ "eval_mean_token_accuracy": 0.8307993857378371,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 98.5061,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 840
+ },
+ {
+ "entropy": 0.5552012033760547,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.5975943207740784,
+ "learning_rate": 6.979947941101646e-05,
+ "loss": 0.5064489841461182,
+ "mean_token_accuracy": 0.8468396887183189,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5844205239515626,
+ "eval_loss": 0.585284948348999,
+ "eval_mean_token_accuracy": 0.8309797065311604,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 98.4982,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 860
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 4.37087437000704e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-880/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-880/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-880/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-880/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-880/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-880/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-880/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-880/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-880/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-880/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-880/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-880/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-880/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-880/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..db9a12c45c9691b2516101959b7e12c1380b9ce4
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-880/trainer_state.json
@@ -0,0 +1,958 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.120627261761158,
+ "eval_steps": 20,
+ "global_step": 880,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ },
+ {
+ "entropy": 0.6141144648194313,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6467089056968689,
+ "learning_rate": 7.078840074092574e-05,
+ "loss": 0.5548617362976074,
+ "mean_token_accuracy": 0.8342376738786698,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.5951976082967908,
+ "eval_loss": 0.5850171446800232,
+ "eval_mean_token_accuracy": 0.8303041870004675,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 98.6601,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 760
+ },
+ {
+ "entropy": 0.5961867064237595,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.41280242800712585,
+ "learning_rate": 7.061001336495465e-05,
+ "loss": 0.5511267185211182,
+ "mean_token_accuracy": 0.8392465204000473,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6167017998320333,
+ "eval_loss": 0.5826791524887085,
+ "eval_mean_token_accuracy": 0.8309678050239434,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 98.6564,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 780
+ },
+ {
+ "entropy": 0.6025628983974457,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.4078947901725769,
+ "learning_rate": 7.042187293157881e-05,
+ "loss": 0.5552026271820069,
+ "mean_token_accuracy": 0.8359778299927711,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.6047008484601974,
+ "eval_loss": 0.5811123847961426,
+ "eval_mean_token_accuracy": 0.8320730804057603,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 98.6033,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 800
+ },
+ {
+ "entropy": 0.6041995719075203,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.543890118598938,
+ "learning_rate": 7.022403268231535e-05,
+ "loss": 0.554384422302246,
+ "mean_token_accuracy": 0.8380729153752327,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.6005127917179901,
+ "eval_loss": 0.5800131559371948,
+ "eval_mean_token_accuracy": 0.8313186547059691,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 98.4119,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 820
+ },
+ {
+ "entropy": 0.567442889396961,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.5015465021133423,
+ "learning_rate": 7.001654860361456e-05,
+ "loss": 0.5208251476287842,
+ "mean_token_accuracy": 0.8442784960453327,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.5785037809878253,
+ "eval_loss": 0.5865591168403625,
+ "eval_mean_token_accuracy": 0.8307993857378371,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 98.5061,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 840
+ },
+ {
+ "entropy": 0.5552012033760547,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.5975943207740784,
+ "learning_rate": 6.979947941101646e-05,
+ "loss": 0.5064489841461182,
+ "mean_token_accuracy": 0.8468396887183189,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5844205239515626,
+ "eval_loss": 0.585284948348999,
+ "eval_mean_token_accuracy": 0.8309797065311604,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 98.4982,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 860
+ },
+ {
+ "entropy": 0.5401638634502888,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.4194582998752594,
+ "learning_rate": 6.957288653253486e-05,
+ "loss": 0.49037814140319824,
+ "mean_token_accuracy": 0.8528496414422989,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.574507936835289,
+ "eval_loss": 0.5836747288703918,
+ "eval_mean_token_accuracy": 0.8325241643391298,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 98.5592,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 880
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 4.47248380551168e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-900/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-900/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-900/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-900/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-900/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-900/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-900/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-900/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-900/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-900/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-900/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-900/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-900/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-900/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..eb4616900b64eebc6b315c1e14355906a0f3042b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-900/trainer_state.json
@@ -0,0 +1,979 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.168878166465621,
+ "eval_steps": 20,
+ "global_step": 900,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ },
+ {
+ "entropy": 0.6141144648194313,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6467089056968689,
+ "learning_rate": 7.078840074092574e-05,
+ "loss": 0.5548617362976074,
+ "mean_token_accuracy": 0.8342376738786698,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.5951976082967908,
+ "eval_loss": 0.5850171446800232,
+ "eval_mean_token_accuracy": 0.8303041870004675,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 98.6601,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 760
+ },
+ {
+ "entropy": 0.5961867064237595,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.41280242800712585,
+ "learning_rate": 7.061001336495465e-05,
+ "loss": 0.5511267185211182,
+ "mean_token_accuracy": 0.8392465204000473,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6167017998320333,
+ "eval_loss": 0.5826791524887085,
+ "eval_mean_token_accuracy": 0.8309678050239434,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 98.6564,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 780
+ },
+ {
+ "entropy": 0.6025628983974457,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.4078947901725769,
+ "learning_rate": 7.042187293157881e-05,
+ "loss": 0.5552026271820069,
+ "mean_token_accuracy": 0.8359778299927711,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.6047008484601974,
+ "eval_loss": 0.5811123847961426,
+ "eval_mean_token_accuracy": 0.8320730804057603,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 98.6033,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 800
+ },
+ {
+ "entropy": 0.6041995719075203,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.543890118598938,
+ "learning_rate": 7.022403268231535e-05,
+ "loss": 0.554384422302246,
+ "mean_token_accuracy": 0.8380729153752327,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.6005127917179901,
+ "eval_loss": 0.5800131559371948,
+ "eval_mean_token_accuracy": 0.8313186547059691,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 98.4119,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 820
+ },
+ {
+ "entropy": 0.567442889396961,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.5015465021133423,
+ "learning_rate": 7.001654860361456e-05,
+ "loss": 0.5208251476287842,
+ "mean_token_accuracy": 0.8442784960453327,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.5785037809878253,
+ "eval_loss": 0.5865591168403625,
+ "eval_mean_token_accuracy": 0.8307993857378371,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 98.5061,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 840
+ },
+ {
+ "entropy": 0.5552012033760547,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.5975943207740784,
+ "learning_rate": 6.979947941101646e-05,
+ "loss": 0.5064489841461182,
+ "mean_token_accuracy": 0.8468396887183189,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5844205239515626,
+ "eval_loss": 0.585284948348999,
+ "eval_mean_token_accuracy": 0.8309797065311604,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 98.4982,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 860
+ },
+ {
+ "entropy": 0.5401638634502888,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.4194582998752594,
+ "learning_rate": 6.957288653253486e-05,
+ "loss": 0.49037814140319824,
+ "mean_token_accuracy": 0.8528496414422989,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.574507936835289,
+ "eval_loss": 0.5836747288703918,
+ "eval_mean_token_accuracy": 0.8325241643391298,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 98.5592,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 880
+ },
+ {
+ "entropy": 0.5606910541653634,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5395355820655823,
+ "learning_rate": 6.933683409127401e-05,
+ "loss": 0.5089571475982666,
+ "mean_token_accuracy": 0.8459660410881042,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5779740023479033,
+ "eval_loss": 0.5816043615341187,
+ "eval_mean_token_accuracy": 0.8326318036304431,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 98.4386,
+ "eval_samples_per_second": 14.425,
+ "eval_steps_per_second": 1.808,
+ "step": 900
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 4.58212871596032e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-920/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-920/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-920/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-920/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-920/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-920/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-920/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-920/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-920/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-920/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-920/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-920/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-920/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-920/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..79cdf0aff34ed73efaa36895f0e33f4a50d1a208
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-920/trainer_state.json
@@ -0,0 +1,1000 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.2171290711700844,
+ "eval_steps": 20,
+ "global_step": 920,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ },
+ {
+ "entropy": 0.6141144648194313,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6467089056968689,
+ "learning_rate": 7.078840074092574e-05,
+ "loss": 0.5548617362976074,
+ "mean_token_accuracy": 0.8342376738786698,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.5951976082967908,
+ "eval_loss": 0.5850171446800232,
+ "eval_mean_token_accuracy": 0.8303041870004675,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 98.6601,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 760
+ },
+ {
+ "entropy": 0.5961867064237595,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.41280242800712585,
+ "learning_rate": 7.061001336495465e-05,
+ "loss": 0.5511267185211182,
+ "mean_token_accuracy": 0.8392465204000473,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6167017998320333,
+ "eval_loss": 0.5826791524887085,
+ "eval_mean_token_accuracy": 0.8309678050239434,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 98.6564,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 780
+ },
+ {
+ "entropy": 0.6025628983974457,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.4078947901725769,
+ "learning_rate": 7.042187293157881e-05,
+ "loss": 0.5552026271820069,
+ "mean_token_accuracy": 0.8359778299927711,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.6047008484601974,
+ "eval_loss": 0.5811123847961426,
+ "eval_mean_token_accuracy": 0.8320730804057603,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 98.6033,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 800
+ },
+ {
+ "entropy": 0.6041995719075203,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.543890118598938,
+ "learning_rate": 7.022403268231535e-05,
+ "loss": 0.554384422302246,
+ "mean_token_accuracy": 0.8380729153752327,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.6005127917179901,
+ "eval_loss": 0.5800131559371948,
+ "eval_mean_token_accuracy": 0.8313186547059691,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 98.4119,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 820
+ },
+ {
+ "entropy": 0.567442889396961,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.5015465021133423,
+ "learning_rate": 7.001654860361456e-05,
+ "loss": 0.5208251476287842,
+ "mean_token_accuracy": 0.8442784960453327,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.5785037809878253,
+ "eval_loss": 0.5865591168403625,
+ "eval_mean_token_accuracy": 0.8307993857378371,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 98.5061,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 840
+ },
+ {
+ "entropy": 0.5552012033760547,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.5975943207740784,
+ "learning_rate": 6.979947941101646e-05,
+ "loss": 0.5064489841461182,
+ "mean_token_accuracy": 0.8468396887183189,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5844205239515626,
+ "eval_loss": 0.585284948348999,
+ "eval_mean_token_accuracy": 0.8309797065311604,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 98.4982,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 860
+ },
+ {
+ "entropy": 0.5401638634502888,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.4194582998752594,
+ "learning_rate": 6.957288653253486e-05,
+ "loss": 0.49037814140319824,
+ "mean_token_accuracy": 0.8528496414422989,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.574507936835289,
+ "eval_loss": 0.5836747288703918,
+ "eval_mean_token_accuracy": 0.8325241643391298,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 98.5592,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 880
+ },
+ {
+ "entropy": 0.5606910541653634,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5395355820655823,
+ "learning_rate": 6.933683409127401e-05,
+ "loss": 0.5089571475982666,
+ "mean_token_accuracy": 0.8459660410881042,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5779740023479033,
+ "eval_loss": 0.5816043615341187,
+ "eval_mean_token_accuracy": 0.8326318036304431,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 98.4386,
+ "eval_samples_per_second": 14.425,
+ "eval_steps_per_second": 1.808,
+ "step": 900
+ },
+ {
+ "entropy": 0.5380903735756875,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.46588417887687683,
+ "learning_rate": 6.909138888728257e-05,
+ "loss": 0.4902662754058838,
+ "mean_token_accuracy": 0.8493925213813782,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.570154261220707,
+ "eval_loss": 0.5827520489692688,
+ "eval_mean_token_accuracy": 0.8321914167216654,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 98.5674,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 920
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 4.68927750014976e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-940/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-940/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-940/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-940/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-940/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-940/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-940/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-940/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-940/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-940/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-940/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-940/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-940/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-940/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..9fd1c0a9b39fd21e5ab06fbaa851b28aa2ded491
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-940/trainer_state.json
@@ -0,0 +1,1021 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.2653799758745476,
+ "eval_steps": 20,
+ "global_step": 940,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ },
+ {
+ "entropy": 0.6141144648194313,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6467089056968689,
+ "learning_rate": 7.078840074092574e-05,
+ "loss": 0.5548617362976074,
+ "mean_token_accuracy": 0.8342376738786698,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.5951976082967908,
+ "eval_loss": 0.5850171446800232,
+ "eval_mean_token_accuracy": 0.8303041870004675,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 98.6601,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 760
+ },
+ {
+ "entropy": 0.5961867064237595,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.41280242800712585,
+ "learning_rate": 7.061001336495465e-05,
+ "loss": 0.5511267185211182,
+ "mean_token_accuracy": 0.8392465204000473,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6167017998320333,
+ "eval_loss": 0.5826791524887085,
+ "eval_mean_token_accuracy": 0.8309678050239434,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 98.6564,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 780
+ },
+ {
+ "entropy": 0.6025628983974457,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.4078947901725769,
+ "learning_rate": 7.042187293157881e-05,
+ "loss": 0.5552026271820069,
+ "mean_token_accuracy": 0.8359778299927711,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.6047008484601974,
+ "eval_loss": 0.5811123847961426,
+ "eval_mean_token_accuracy": 0.8320730804057603,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 98.6033,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 800
+ },
+ {
+ "entropy": 0.6041995719075203,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.543890118598938,
+ "learning_rate": 7.022403268231535e-05,
+ "loss": 0.554384422302246,
+ "mean_token_accuracy": 0.8380729153752327,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.6005127917179901,
+ "eval_loss": 0.5800131559371948,
+ "eval_mean_token_accuracy": 0.8313186547059691,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 98.4119,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 820
+ },
+ {
+ "entropy": 0.567442889396961,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.5015465021133423,
+ "learning_rate": 7.001654860361456e-05,
+ "loss": 0.5208251476287842,
+ "mean_token_accuracy": 0.8442784960453327,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.5785037809878253,
+ "eval_loss": 0.5865591168403625,
+ "eval_mean_token_accuracy": 0.8307993857378371,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 98.5061,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 840
+ },
+ {
+ "entropy": 0.5552012033760547,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.5975943207740784,
+ "learning_rate": 6.979947941101646e-05,
+ "loss": 0.5064489841461182,
+ "mean_token_accuracy": 0.8468396887183189,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5844205239515626,
+ "eval_loss": 0.585284948348999,
+ "eval_mean_token_accuracy": 0.8309797065311604,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 98.4982,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 860
+ },
+ {
+ "entropy": 0.5401638634502888,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.4194582998752594,
+ "learning_rate": 6.957288653253486e-05,
+ "loss": 0.49037814140319824,
+ "mean_token_accuracy": 0.8528496414422989,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.574507936835289,
+ "eval_loss": 0.5836747288703918,
+ "eval_mean_token_accuracy": 0.8325241643391298,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 98.5592,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 880
+ },
+ {
+ "entropy": 0.5606910541653634,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5395355820655823,
+ "learning_rate": 6.933683409127401e-05,
+ "loss": 0.5089571475982666,
+ "mean_token_accuracy": 0.8459660410881042,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5779740023479033,
+ "eval_loss": 0.5816043615341187,
+ "eval_mean_token_accuracy": 0.8326318036304431,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 98.4386,
+ "eval_samples_per_second": 14.425,
+ "eval_steps_per_second": 1.808,
+ "step": 900
+ },
+ {
+ "entropy": 0.5380903735756875,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.46588417887687683,
+ "learning_rate": 6.909138888728257e-05,
+ "loss": 0.4902662754058838,
+ "mean_token_accuracy": 0.8493925213813782,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.570154261220707,
+ "eval_loss": 0.5827520489692688,
+ "eval_mean_token_accuracy": 0.8321914167216654,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 98.5674,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 920
+ },
+ {
+ "entropy": 0.558555257320404,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.44576048851013184,
+ "learning_rate": 6.883662037864996e-05,
+ "loss": 0.5000699520111084,
+ "mean_token_accuracy": 0.8475695163011551,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5715417749761196,
+ "eval_loss": 0.5831062197685242,
+ "eval_mean_token_accuracy": 0.8327409899636601,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 98.568,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 940
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 4.781791461888e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-960/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-960/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-960/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-960/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-960/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-960/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-960/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-960/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-960/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-960/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-960/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-960/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-960/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-960/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..f44d7b85b02c2370bf32de50e5113147774fef16
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-960/trainer_state.json
@@ -0,0 +1,1042 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.313630880579011,
+ "eval_steps": 20,
+ "global_step": 960,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ },
+ {
+ "entropy": 0.6141144648194313,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6467089056968689,
+ "learning_rate": 7.078840074092574e-05,
+ "loss": 0.5548617362976074,
+ "mean_token_accuracy": 0.8342376738786698,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.5951976082967908,
+ "eval_loss": 0.5850171446800232,
+ "eval_mean_token_accuracy": 0.8303041870004675,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 98.6601,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 760
+ },
+ {
+ "entropy": 0.5961867064237595,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.41280242800712585,
+ "learning_rate": 7.061001336495465e-05,
+ "loss": 0.5511267185211182,
+ "mean_token_accuracy": 0.8392465204000473,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6167017998320333,
+ "eval_loss": 0.5826791524887085,
+ "eval_mean_token_accuracy": 0.8309678050239434,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 98.6564,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 780
+ },
+ {
+ "entropy": 0.6025628983974457,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.4078947901725769,
+ "learning_rate": 7.042187293157881e-05,
+ "loss": 0.5552026271820069,
+ "mean_token_accuracy": 0.8359778299927711,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.6047008484601974,
+ "eval_loss": 0.5811123847961426,
+ "eval_mean_token_accuracy": 0.8320730804057603,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 98.6033,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 800
+ },
+ {
+ "entropy": 0.6041995719075203,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.543890118598938,
+ "learning_rate": 7.022403268231535e-05,
+ "loss": 0.554384422302246,
+ "mean_token_accuracy": 0.8380729153752327,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.6005127917179901,
+ "eval_loss": 0.5800131559371948,
+ "eval_mean_token_accuracy": 0.8313186547059691,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 98.4119,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 820
+ },
+ {
+ "entropy": 0.567442889396961,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.5015465021133423,
+ "learning_rate": 7.001654860361456e-05,
+ "loss": 0.5208251476287842,
+ "mean_token_accuracy": 0.8442784960453327,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.5785037809878253,
+ "eval_loss": 0.5865591168403625,
+ "eval_mean_token_accuracy": 0.8307993857378371,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 98.5061,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 840
+ },
+ {
+ "entropy": 0.5552012033760547,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.5975943207740784,
+ "learning_rate": 6.979947941101646e-05,
+ "loss": 0.5064489841461182,
+ "mean_token_accuracy": 0.8468396887183189,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5844205239515626,
+ "eval_loss": 0.585284948348999,
+ "eval_mean_token_accuracy": 0.8309797065311604,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 98.4982,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 860
+ },
+ {
+ "entropy": 0.5401638634502888,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.4194582998752594,
+ "learning_rate": 6.957288653253486e-05,
+ "loss": 0.49037814140319824,
+ "mean_token_accuracy": 0.8528496414422989,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.574507936835289,
+ "eval_loss": 0.5836747288703918,
+ "eval_mean_token_accuracy": 0.8325241643391298,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 98.5592,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 880
+ },
+ {
+ "entropy": 0.5606910541653634,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5395355820655823,
+ "learning_rate": 6.933683409127401e-05,
+ "loss": 0.5089571475982666,
+ "mean_token_accuracy": 0.8459660410881042,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5779740023479033,
+ "eval_loss": 0.5816043615341187,
+ "eval_mean_token_accuracy": 0.8326318036304431,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 98.4386,
+ "eval_samples_per_second": 14.425,
+ "eval_steps_per_second": 1.808,
+ "step": 900
+ },
+ {
+ "entropy": 0.5380903735756875,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.46588417887687683,
+ "learning_rate": 6.909138888728257e-05,
+ "loss": 0.4902662754058838,
+ "mean_token_accuracy": 0.8493925213813782,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.570154261220707,
+ "eval_loss": 0.5827520489692688,
+ "eval_mean_token_accuracy": 0.8321914167216654,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 98.5674,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 920
+ },
+ {
+ "entropy": 0.558555257320404,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.44576048851013184,
+ "learning_rate": 6.883662037864996e-05,
+ "loss": 0.5000699520111084,
+ "mean_token_accuracy": 0.8475695163011551,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5715417749761196,
+ "eval_loss": 0.5831062197685242,
+ "eval_mean_token_accuracy": 0.8327409899636601,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 98.568,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 940
+ },
+ {
+ "entropy": 0.5582522705197335,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.38966405391693115,
+ "learning_rate": 6.857260066185045e-05,
+ "loss": 0.5132341861724854,
+ "mean_token_accuracy": 0.8466864034533501,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5823918579334624,
+ "eval_loss": 0.5797588229179382,
+ "eval_mean_token_accuracy": 0.8326037566313583,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 98.6222,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 960
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 4.87588516380672e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-980/README.md b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-980/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..6b69fa97fbe239b1f2988d3e366763603689ad4b
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-980/README.md
@@ -0,0 +1,209 @@
+---
+base_model: Qwen/Qwen3-14B-Base
+library_name: peft
+pipeline_tag: text-generation
+tags:
+- base_model:adapter:Qwen/Qwen3-14B-Base
+- lora
+- sft
+- transformers
+- trl
+---
+
+# Model Card for Model ID
+
+
+
+
+
+## Model Details
+
+### Model Description
+
+
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+
+
+### Direct Use
+
+
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+
+
+[More Information Needed]
+
+### Recommendations
+
+
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+
+
+[More Information Needed]
+
+### Training Procedure
+
+
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed]
+
+#### Speeds, Sizes, Times [optional]
+
+
+
+[More Information Needed]
+
+## Evaluation
+
+
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+
+
+[More Information Needed]
+
+#### Factors
+
+
+
+[More Information Needed]
+
+#### Metrics
+
+
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+
+
+[More Information Needed]
+
+## Environmental Impact
+
+
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.18.1
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-980/adapter_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-980/adapter_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..7ccfc07021d124c37907d00e2fac38d632c0b509
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-980/adapter_config.json
@@ -0,0 +1,46 @@
+{
+ "alora_invocation_tokens": null,
+ "alpha_pattern": {},
+ "arrow_config": null,
+ "auto_mapping": null,
+ "base_model_name_or_path": "Qwen/Qwen3-14B-Base",
+ "bias": "none",
+ "corda_config": null,
+ "ensure_weight_tying": false,
+ "eva_config": null,
+ "exclude_modules": null,
+ "fan_in_fan_out": false,
+ "inference_mode": true,
+ "init_lora_weights": true,
+ "layer_replication": null,
+ "layers_pattern": null,
+ "layers_to_transform": null,
+ "loftq_config": {},
+ "lora_alpha": 64,
+ "lora_bias": false,
+ "lora_dropout": 0.009940685857937281,
+ "megatron_config": null,
+ "megatron_core": "megatron.core",
+ "modules_to_save": null,
+ "peft_type": "LORA",
+ "peft_version": "0.18.1",
+ "qalora_group_size": 16,
+ "r": 64,
+ "rank_pattern": {},
+ "revision": null,
+ "target_modules": [
+ "gate_proj",
+ "q_proj",
+ "down_proj",
+ "o_proj",
+ "k_proj",
+ "v_proj",
+ "up_proj"
+ ],
+ "target_parameters": null,
+ "task_type": "CAUSAL_LM",
+ "trainable_token_indices": null,
+ "use_dora": false,
+ "use_qalora": false,
+ "use_rslora": false
+}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-980/chat_template.jinja b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-980/chat_template.jinja
new file mode 100644
index 0000000000000000000000000000000000000000..699ff8df401fe4788525e9c1f9b86a99eadd6230
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-980/chat_template.jinja
@@ -0,0 +1,85 @@
+{%- if tools %}
+ {{- '<|im_start|>system\n' }}
+ {%- if messages[0].role == 'system' %}
+ {{- messages[0].content + '\n\n' }}
+ {%- endif %}
+ {{- "# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within XML tags:\n" }}
+ {%- for tool in tools %}
+ {{- "\n" }}
+ {{- tool | tojson }}
+ {%- endfor %}
+ {{- "\n\n\nFor each function call, return a json object with function name and arguments within XML tags:\n\n{\"name\": , \"arguments\": }\n<|im_end|>\n" }}
+{%- else %}
+ {%- if messages[0].role == 'system' %}
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
+ {%- endif %}
+{%- endif %}
+{%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
+{%- for message in messages[::-1] %}
+ {%- set index = (messages|length - 1) - loop.index0 %}
+ {%- if ns.multi_step_tool and message.role == "user" and not(message.content.startswith('') and message.content.endswith('')) %}
+ {%- set ns.multi_step_tool = false %}
+ {%- set ns.last_query_index = index %}
+ {%- endif %}
+{%- endfor %}
+{%- for message in messages %}
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+ {%- elif message.role == "assistant" %}
+ {%- set content = message.content %}
+ {%- set reasoning_content = '' %}
+ {%- if message.reasoning_content is defined and message.reasoning_content is not none %}
+ {%- set reasoning_content = message.reasoning_content %}
+ {%- else %}
+ {%- if '' in message.content %}
+ {%- set content = message.content.split('')[-1].lstrip('\n') %}
+ {%- set reasoning_content = message.content.split('')[0].rstrip('\n').split('')[-1].lstrip('\n') %}
+ {%- endif %}
+ {%- endif %}
+ {%- if loop.index0 > ns.last_query_index %}
+ {%- if loop.last or (not loop.last and reasoning_content) %}
+ {{- '<|im_start|>' + message.role + '\n\n' + reasoning_content.strip('\n') + '\n\n\n' + content.lstrip('\n') }}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- else %}
+ {{- '<|im_start|>' + message.role + '\n' + content }}
+ {%- endif %}
+ {%- if message.tool_calls %}
+ {%- for tool_call in message.tool_calls %}
+ {%- if (loop.first and content) or (not loop.first) %}
+ {{- '\n' }}
+ {%- endif %}
+ {%- if tool_call.function %}
+ {%- set tool_call = tool_call.function %}
+ {%- endif %}
+ {{- '\n{"name": "' }}
+ {{- tool_call.name }}
+ {{- '", "arguments": ' }}
+ {%- if tool_call.arguments is string %}
+ {{- tool_call.arguments }}
+ {%- else %}
+ {{- tool_call.arguments | tojson }}
+ {%- endif %}
+ {{- '}\n' }}
+ {%- endfor %}
+ {%- endif %}
+ {{- '<|im_end|>\n' }}
+ {%- elif message.role == "tool" %}
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
+ {{- '<|im_start|>user' }}
+ {%- endif %}
+ {{- '\n\n' }}
+ {{- message.content }}
+ {{- '\n' }}
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+ {{- '<|im_end|>\n' }}
+ {%- endif %}
+ {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+ {{- '<|im_start|>assistant\n' }}
+ {%- if enable_thinking is defined and enable_thinking is false %}
+ {{- '\n\n\n\n' }}
+ {%- endif %}
+{%- endif %}
\ No newline at end of file
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-980/tokenizer_config.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-980/tokenizer_config.json
new file mode 100644
index 0000000000000000000000000000000000000000..c960ecf0d33fd7b8c99d12680c0e74a82b36d446
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-980/tokenizer_config.json
@@ -0,0 +1,29 @@
+{
+ "add_prefix_space": false,
+ "backend": "tokenizers",
+ "bos_token": null,
+ "clean_up_tokenization_spaces": false,
+ "eos_token": "<|endoftext|>",
+ "errors": "replace",
+ "extra_special_tokens": [
+ "<|im_start|>",
+ "<|im_end|>",
+ "<|object_ref_start|>",
+ "<|object_ref_end|>",
+ "<|box_start|>",
+ "<|box_end|>",
+ "<|quad_start|>",
+ "<|quad_end|>",
+ "<|vision_start|>",
+ "<|vision_end|>",
+ "<|vision_pad|>",
+ "<|image_pad|>",
+ "<|video_pad|>"
+ ],
+ "is_local": false,
+ "model_max_length": 131072,
+ "pad_token": "<|endoftext|>",
+ "split_special_tokens": false,
+ "tokenizer_class": "Qwen2Tokenizer",
+ "unk_token": null
+}
diff --git a/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-980/trainer_state.json b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-980/trainer_state.json
new file mode 100644
index 0000000000000000000000000000000000000000..daca50de5d52a815bebf47bb9ff873e966083453
--- /dev/null
+++ b/overgeneralisation_original_Estonian/Qwen3-14B-Base_overgeneralisation_splits_original_features_train_overgeneralisation_splits_original_features_test2/checkpoint-980/trainer_state.json
@@ -0,0 +1,1063 @@
+{
+ "best_global_step": null,
+ "best_metric": null,
+ "best_model_checkpoint": null,
+ "epoch": 2.361881785283474,
+ "eval_steps": 20,
+ "global_step": 980,
+ "is_hyper_param_search": false,
+ "is_local_process_zero": true,
+ "is_world_process_zero": true,
+ "log_history": [
+ {
+ "entropy": 1.8880569130182265,
+ "epoch": 0.04825090470446321,
+ "grad_norm": 0.5548511743545532,
+ "learning_rate": 3.3097059104127273e-06,
+ "loss": 1.863978958129883,
+ "mean_token_accuracy": 0.621908786892891,
+ "num_tokens": 50237.0,
+ "step": 20
+ },
+ {
+ "epoch": 0.04825090470446321,
+ "eval_entropy": 1.8981187470843284,
+ "eval_loss": 1.9387791156768799,
+ "eval_mean_token_accuracy": 0.6191755044326354,
+ "eval_num_tokens": 50237.0,
+ "eval_runtime": 99.1278,
+ "eval_samples_per_second": 14.325,
+ "eval_steps_per_second": 1.796,
+ "step": 20
+ },
+ {
+ "entropy": 1.845469006896019,
+ "epoch": 0.09650180940892641,
+ "grad_norm": 0.3859543204307556,
+ "learning_rate": 6.793606868741914e-06,
+ "loss": 1.8429210662841797,
+ "mean_token_accuracy": 0.6286612510681152,
+ "num_tokens": 101098.0,
+ "step": 40
+ },
+ {
+ "epoch": 0.09650180940892641,
+ "eval_entropy": 1.8266146584842982,
+ "eval_loss": 1.8124289512634277,
+ "eval_mean_token_accuracy": 0.636085176400924,
+ "eval_num_tokens": 101098.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 40
+ },
+ {
+ "entropy": 1.6761731654405594,
+ "epoch": 0.14475271411338964,
+ "grad_norm": 0.763229489326477,
+ "learning_rate": 1.0277507827071102e-05,
+ "loss": 1.59182710647583,
+ "mean_token_accuracy": 0.6615026861429214,
+ "num_tokens": 150385.0,
+ "step": 60
+ },
+ {
+ "epoch": 0.14475271411338964,
+ "eval_entropy": 1.448009472884489,
+ "eval_loss": 1.4045010805130005,
+ "eval_mean_token_accuracy": 0.6956760652279586,
+ "eval_num_tokens": 150385.0,
+ "eval_runtime": 98.6634,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 60
+ },
+ {
+ "entropy": 1.2918378412723541,
+ "epoch": 0.19300361881785283,
+ "grad_norm": 0.6131361126899719,
+ "learning_rate": 1.3761408785400288e-05,
+ "loss": 1.2077620506286622,
+ "mean_token_accuracy": 0.715845774114132,
+ "num_tokens": 197354.0,
+ "step": 80
+ },
+ {
+ "epoch": 0.19300361881785283,
+ "eval_entropy": 1.0959123356288738,
+ "eval_loss": 1.025644063949585,
+ "eval_mean_token_accuracy": 0.7453202035989654,
+ "eval_num_tokens": 197354.0,
+ "eval_runtime": 98.6386,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 80
+ },
+ {
+ "entropy": 0.9881208389997482,
+ "epoch": 0.24125452352231605,
+ "grad_norm": 0.4983534514904022,
+ "learning_rate": 1.7245309743729473e-05,
+ "loss": 0.9109245300292969,
+ "mean_token_accuracy": 0.7622973442077636,
+ "num_tokens": 248557.0,
+ "step": 100
+ },
+ {
+ "epoch": 0.24125452352231605,
+ "eval_entropy": 0.9157560136211053,
+ "eval_loss": 0.8661546111106873,
+ "eval_mean_token_accuracy": 0.7745042727904373,
+ "eval_num_tokens": 248557.0,
+ "eval_runtime": 98.5282,
+ "eval_samples_per_second": 14.412,
+ "eval_steps_per_second": 1.807,
+ "step": 100
+ },
+ {
+ "entropy": 0.8518766984343529,
+ "epoch": 0.28950542822677927,
+ "grad_norm": 0.635079026222229,
+ "learning_rate": 2.0729210702058663e-05,
+ "loss": 0.8058656692504883,
+ "mean_token_accuracy": 0.7847469851374627,
+ "num_tokens": 298041.0,
+ "step": 120
+ },
+ {
+ "epoch": 0.28950542822677927,
+ "eval_entropy": 0.838603059227547,
+ "eval_loss": 0.7992516756057739,
+ "eval_mean_token_accuracy": 0.7849764954508021,
+ "eval_num_tokens": 298041.0,
+ "eval_runtime": 98.5627,
+ "eval_samples_per_second": 14.407,
+ "eval_steps_per_second": 1.806,
+ "step": 120
+ },
+ {
+ "entropy": 0.8069365546107292,
+ "epoch": 0.33775633293124246,
+ "grad_norm": 0.5280472636222839,
+ "learning_rate": 2.421311166038785e-05,
+ "loss": 0.7562496662139893,
+ "mean_token_accuracy": 0.7897997036576271,
+ "num_tokens": 348955.0,
+ "step": 140
+ },
+ {
+ "epoch": 0.33775633293124246,
+ "eval_entropy": 0.8161682168419442,
+ "eval_loss": 0.7599642276763916,
+ "eval_mean_token_accuracy": 0.7912511644738444,
+ "eval_num_tokens": 348955.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 140
+ },
+ {
+ "entropy": 0.7952415809035301,
+ "epoch": 0.38600723763570566,
+ "grad_norm": 0.5387083888053894,
+ "learning_rate": 2.7697012618717034e-05,
+ "loss": 0.740808629989624,
+ "mean_token_accuracy": 0.7941806614398956,
+ "num_tokens": 395798.0,
+ "step": 160
+ },
+ {
+ "epoch": 0.38600723763570566,
+ "eval_entropy": 0.7856516191798649,
+ "eval_loss": 0.7324908971786499,
+ "eval_mean_token_accuracy": 0.7980317825011993,
+ "eval_num_tokens": 395798.0,
+ "eval_runtime": 98.6378,
+ "eval_samples_per_second": 14.396,
+ "eval_steps_per_second": 1.805,
+ "step": 160
+ },
+ {
+ "entropy": 0.7537026956677437,
+ "epoch": 0.43425814234016885,
+ "grad_norm": 0.5521425604820251,
+ "learning_rate": 3.118091357704622e-05,
+ "loss": 0.7073616027832031,
+ "mean_token_accuracy": 0.8006261065602303,
+ "num_tokens": 446437.0,
+ "step": 180
+ },
+ {
+ "epoch": 0.43425814234016885,
+ "eval_entropy": 0.7470693678668375,
+ "eval_loss": 0.711143434047699,
+ "eval_mean_token_accuracy": 0.801459076029531,
+ "eval_num_tokens": 446437.0,
+ "eval_runtime": 98.6667,
+ "eval_samples_per_second": 14.392,
+ "eval_steps_per_second": 1.804,
+ "step": 180
+ },
+ {
+ "entropy": 0.7328549668192863,
+ "epoch": 0.4825090470446321,
+ "grad_norm": 0.4901895821094513,
+ "learning_rate": 3.4664814535375406e-05,
+ "loss": 0.6874266624450683,
+ "mean_token_accuracy": 0.8066758885979652,
+ "num_tokens": 497682.0,
+ "step": 200
+ },
+ {
+ "epoch": 0.4825090470446321,
+ "eval_entropy": 0.7374835154983435,
+ "eval_loss": 0.6967977285385132,
+ "eval_mean_token_accuracy": 0.8055616520094068,
+ "eval_num_tokens": 497682.0,
+ "eval_runtime": 98.6234,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 200
+ },
+ {
+ "entropy": 0.7372006759047508,
+ "epoch": 0.5307599517490953,
+ "grad_norm": 0.5516473650932312,
+ "learning_rate": 3.814871549370459e-05,
+ "loss": 0.6906380653381348,
+ "mean_token_accuracy": 0.8057604253292083,
+ "num_tokens": 544670.0,
+ "step": 220
+ },
+ {
+ "epoch": 0.5307599517490953,
+ "eval_entropy": 0.710068748573239,
+ "eval_loss": 0.6860950589179993,
+ "eval_mean_token_accuracy": 0.8069355440943429,
+ "eval_num_tokens": 544670.0,
+ "eval_runtime": 98.4618,
+ "eval_samples_per_second": 14.422,
+ "eval_steps_per_second": 1.808,
+ "step": 220
+ },
+ {
+ "entropy": 0.7099859565496445,
+ "epoch": 0.5790108564535585,
+ "grad_norm": 0.457378625869751,
+ "learning_rate": 4.1632616452033785e-05,
+ "loss": 0.6674468517303467,
+ "mean_token_accuracy": 0.8087703332304954,
+ "num_tokens": 592926.0,
+ "step": 240
+ },
+ {
+ "epoch": 0.5790108564535585,
+ "eval_entropy": 0.717441116156203,
+ "eval_loss": 0.6765118837356567,
+ "eval_mean_token_accuracy": 0.8095786246690857,
+ "eval_num_tokens": 592926.0,
+ "eval_runtime": 98.6828,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 240
+ },
+ {
+ "entropy": 0.7170855522155761,
+ "epoch": 0.6272617611580217,
+ "grad_norm": 0.4604854881763458,
+ "learning_rate": 4.5116517410362974e-05,
+ "loss": 0.6647133827209473,
+ "mean_token_accuracy": 0.8115006521344185,
+ "num_tokens": 639077.0,
+ "step": 260
+ },
+ {
+ "epoch": 0.6272617611580217,
+ "eval_entropy": 0.7082702219486237,
+ "eval_loss": 0.6722468733787537,
+ "eval_mean_token_accuracy": 0.8100237367528208,
+ "eval_num_tokens": 639077.0,
+ "eval_runtime": 98.6242,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 260
+ },
+ {
+ "entropy": 0.691286925971508,
+ "epoch": 0.6755126658624849,
+ "grad_norm": 0.4626481831073761,
+ "learning_rate": 4.860041836869215e-05,
+ "loss": 0.6543298721313476,
+ "mean_token_accuracy": 0.8142751976847649,
+ "num_tokens": 691028.0,
+ "step": 280
+ },
+ {
+ "epoch": 0.6755126658624849,
+ "eval_entropy": 0.7341568747932992,
+ "eval_loss": 0.658997654914856,
+ "eval_mean_token_accuracy": 0.813439255684949,
+ "eval_num_tokens": 691028.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 280
+ },
+ {
+ "entropy": 0.691513468325138,
+ "epoch": 0.7237635705669482,
+ "grad_norm": 0.5434145331382751,
+ "learning_rate": 5.2084319327021346e-05,
+ "loss": 0.6393396377563476,
+ "mean_token_accuracy": 0.8172704800963402,
+ "num_tokens": 740255.0,
+ "step": 300
+ },
+ {
+ "epoch": 0.7237635705669482,
+ "eval_entropy": 0.699160492487168,
+ "eval_loss": 0.6522852182388306,
+ "eval_mean_token_accuracy": 0.8143363323774231,
+ "eval_num_tokens": 740255.0,
+ "eval_runtime": 98.6233,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 300
+ },
+ {
+ "entropy": 0.6918429866433143,
+ "epoch": 0.7720144752714113,
+ "grad_norm": 0.47131890058517456,
+ "learning_rate": 5.556822028535052e-05,
+ "loss": 0.6503581523895263,
+ "mean_token_accuracy": 0.8149650946259499,
+ "num_tokens": 789529.0,
+ "step": 320
+ },
+ {
+ "epoch": 0.7720144752714113,
+ "eval_entropy": 0.6949931701247611,
+ "eval_loss": 0.6464414596557617,
+ "eval_mean_token_accuracy": 0.816310279489903,
+ "eval_num_tokens": 789529.0,
+ "eval_runtime": 98.4759,
+ "eval_samples_per_second": 14.42,
+ "eval_steps_per_second": 1.808,
+ "step": 320
+ },
+ {
+ "entropy": 0.6926439955830574,
+ "epoch": 0.8202653799758746,
+ "grad_norm": 0.46790942549705505,
+ "learning_rate": 5.905212124367972e-05,
+ "loss": 0.6453651428222656,
+ "mean_token_accuracy": 0.8171535924077034,
+ "num_tokens": 841584.0,
+ "step": 340
+ },
+ {
+ "epoch": 0.8202653799758746,
+ "eval_entropy": 0.6810062698433909,
+ "eval_loss": 0.6440736651420593,
+ "eval_mean_token_accuracy": 0.8166696479481258,
+ "eval_num_tokens": 841584.0,
+ "eval_runtime": 98.6479,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 340
+ },
+ {
+ "entropy": 0.7038845136761666,
+ "epoch": 0.8685162846803377,
+ "grad_norm": 0.5038695335388184,
+ "learning_rate": 6.253602220200891e-05,
+ "loss": 0.6558614253997803,
+ "mean_token_accuracy": 0.8127013072371483,
+ "num_tokens": 885769.0,
+ "step": 360
+ },
+ {
+ "epoch": 0.8685162846803377,
+ "eval_entropy": 0.6752923710292644,
+ "eval_loss": 0.6410524845123291,
+ "eval_mean_token_accuracy": 0.817534717950928,
+ "eval_num_tokens": 885769.0,
+ "eval_runtime": 98.5986,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 360
+ },
+ {
+ "entropy": 0.6897840186953544,
+ "epoch": 0.916767189384801,
+ "grad_norm": 0.5168097615242004,
+ "learning_rate": 6.601992316033808e-05,
+ "loss": 0.6328047752380371,
+ "mean_token_accuracy": 0.8181755512952804,
+ "num_tokens": 930056.0,
+ "step": 380
+ },
+ {
+ "epoch": 0.916767189384801,
+ "eval_entropy": 0.6816390936964014,
+ "eval_loss": 0.6402509212493896,
+ "eval_mean_token_accuracy": 0.8172605586855599,
+ "eval_num_tokens": 930056.0,
+ "eval_runtime": 98.4787,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 380
+ },
+ {
+ "entropy": 0.6798740655183793,
+ "epoch": 0.9650180940892642,
+ "grad_norm": 0.4609147906303406,
+ "learning_rate": 6.950382411866727e-05,
+ "loss": 0.6358067989349365,
+ "mean_token_accuracy": 0.8193414479494094,
+ "num_tokens": 976892.0,
+ "step": 400
+ },
+ {
+ "epoch": 0.9650180940892642,
+ "eval_entropy": 0.6841439577970612,
+ "eval_loss": 0.6335230469703674,
+ "eval_mean_token_accuracy": 0.8190797965848045,
+ "eval_num_tokens": 976892.0,
+ "eval_runtime": 98.5983,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 400
+ },
+ {
+ "entropy": 0.6647876103719076,
+ "epoch": 1.0120627261761157,
+ "grad_norm": 0.46518462896347046,
+ "learning_rate": 7.229074030587624e-05,
+ "loss": 0.6112009525299072,
+ "mean_token_accuracy": 0.8212849283829714,
+ "num_tokens": 1024472.0,
+ "step": 420
+ },
+ {
+ "epoch": 1.0120627261761157,
+ "eval_entropy": 0.6612423688508151,
+ "eval_loss": 0.6311497688293457,
+ "eval_mean_token_accuracy": 0.8194578638237514,
+ "eval_num_tokens": 1024472.0,
+ "eval_runtime": 98.4872,
+ "eval_samples_per_second": 14.418,
+ "eval_steps_per_second": 1.807,
+ "step": 420
+ },
+ {
+ "entropy": 0.6288895070552826,
+ "epoch": 1.060313630880579,
+ "grad_norm": 0.4342021644115448,
+ "learning_rate": 7.228358026812833e-05,
+ "loss": 0.5853999614715576,
+ "mean_token_accuracy": 0.82901301831007,
+ "num_tokens": 1076964.0,
+ "step": 440
+ },
+ {
+ "epoch": 1.060313630880579,
+ "eval_entropy": 0.6567474329069759,
+ "eval_loss": 0.6245927810668945,
+ "eval_mean_token_accuracy": 0.8211770050981072,
+ "eval_num_tokens": 1076964.0,
+ "eval_runtime": 98.6093,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 440
+ },
+ {
+ "entropy": 0.639747454226017,
+ "epoch": 1.1085645355850422,
+ "grad_norm": 0.5338298082351685,
+ "learning_rate": 7.226619357333845e-05,
+ "loss": 0.5847910404205322,
+ "mean_token_accuracy": 0.8279331669211387,
+ "num_tokens": 1126064.0,
+ "step": 460
+ },
+ {
+ "epoch": 1.1085645355850422,
+ "eval_entropy": 0.6399250211340658,
+ "eval_loss": 0.6245712637901306,
+ "eval_mean_token_accuracy": 0.8209733875949731,
+ "eval_num_tokens": 1126064.0,
+ "eval_runtime": 98.6581,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 460
+ },
+ {
+ "entropy": 0.6248803421854973,
+ "epoch": 1.1568154402895054,
+ "grad_norm": 0.4402505159378052,
+ "learning_rate": 7.223858514173552e-05,
+ "loss": 0.5847790241241455,
+ "mean_token_accuracy": 0.8273933246731758,
+ "num_tokens": 1176912.0,
+ "step": 480
+ },
+ {
+ "epoch": 1.1568154402895054,
+ "eval_entropy": 0.6459465750147787,
+ "eval_loss": 0.615716278553009,
+ "eval_mean_token_accuracy": 0.8228987813665626,
+ "eval_num_tokens": 1176912.0,
+ "eval_runtime": 98.5987,
+ "eval_samples_per_second": 14.402,
+ "eval_steps_per_second": 1.805,
+ "step": 480
+ },
+ {
+ "entropy": 0.6356995061039925,
+ "epoch": 1.2050663449939687,
+ "grad_norm": 0.5318694114685059,
+ "learning_rate": 7.220076278617914e-05,
+ "loss": 0.5902805328369141,
+ "mean_token_accuracy": 0.8279137581586837,
+ "num_tokens": 1223383.0,
+ "step": 500
+ },
+ {
+ "epoch": 1.2050663449939687,
+ "eval_entropy": 0.6509443488013879,
+ "eval_loss": 0.6174858808517456,
+ "eval_mean_token_accuracy": 0.8229377031996009,
+ "eval_num_tokens": 1223383.0,
+ "eval_runtime": 98.6156,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 500
+ },
+ {
+ "entropy": 0.6378119193017483,
+ "epoch": 1.2533172496984317,
+ "grad_norm": 0.5385730266571045,
+ "learning_rate": 7.215273720994853e-05,
+ "loss": 0.5850183486938476,
+ "mean_token_accuracy": 0.8282771646976471,
+ "num_tokens": 1273534.0,
+ "step": 520
+ },
+ {
+ "epoch": 1.2533172496984317,
+ "eval_entropy": 0.6438969477508845,
+ "eval_loss": 0.6134508848190308,
+ "eval_mean_token_accuracy": 0.823984161186754,
+ "eval_num_tokens": 1273534.0,
+ "eval_runtime": 98.4814,
+ "eval_samples_per_second": 14.419,
+ "eval_steps_per_second": 1.807,
+ "step": 520
+ },
+ {
+ "entropy": 0.6330174207687378,
+ "epoch": 1.301568154402895,
+ "grad_norm": 0.4950766861438751,
+ "learning_rate": 7.209452200371367e-05,
+ "loss": 0.5886627674102783,
+ "mean_token_accuracy": 0.828557687997818,
+ "num_tokens": 1323351.0,
+ "step": 540
+ },
+ {
+ "epoch": 1.301568154402895,
+ "eval_entropy": 0.6342737025089478,
+ "eval_loss": 0.611145555973053,
+ "eval_mean_token_accuracy": 0.8250270244780551,
+ "eval_num_tokens": 1323351.0,
+ "eval_runtime": 98.6698,
+ "eval_samples_per_second": 14.391,
+ "eval_steps_per_second": 1.804,
+ "step": 540
+ },
+ {
+ "entropy": 0.6270076125860214,
+ "epoch": 1.3498190591073582,
+ "grad_norm": 0.435604989528656,
+ "learning_rate": 7.20261336416893e-05,
+ "loss": 0.5834732532501221,
+ "mean_token_accuracy": 0.8290600389242172,
+ "num_tokens": 1371508.0,
+ "step": 560
+ },
+ {
+ "epoch": 1.3498190591073582,
+ "eval_entropy": 0.6624904408883513,
+ "eval_loss": 0.6063718795776367,
+ "eval_mean_token_accuracy": 0.825260305002834,
+ "eval_num_tokens": 1371508.0,
+ "eval_runtime": 98.6348,
+ "eval_samples_per_second": 14.397,
+ "eval_steps_per_second": 1.805,
+ "step": 560
+ },
+ {
+ "entropy": 0.6258987173438072,
+ "epoch": 1.3980699638118215,
+ "grad_norm": 0.45945441722869873,
+ "learning_rate": 7.194759147697293e-05,
+ "loss": 0.5705549716949463,
+ "mean_token_accuracy": 0.8329991281032563,
+ "num_tokens": 1415221.0,
+ "step": 580
+ },
+ {
+ "epoch": 1.3980699638118215,
+ "eval_entropy": 0.6500632069753797,
+ "eval_loss": 0.6059838533401489,
+ "eval_mean_token_accuracy": 0.8262855604793249,
+ "eval_num_tokens": 1415221.0,
+ "eval_runtime": 98.6469,
+ "eval_samples_per_second": 14.395,
+ "eval_steps_per_second": 1.804,
+ "step": 580
+ },
+ {
+ "entropy": 0.6301489062607288,
+ "epoch": 1.4463208685162847,
+ "grad_norm": 0.5217350721359253,
+ "learning_rate": 7.18589177360681e-05,
+ "loss": 0.5778427124023438,
+ "mean_token_accuracy": 0.8305584967136384,
+ "num_tokens": 1463356.0,
+ "step": 600
+ },
+ {
+ "epoch": 1.4463208685162847,
+ "eval_entropy": 0.6478241879618569,
+ "eval_loss": 0.6014655828475952,
+ "eval_mean_token_accuracy": 0.8265940369514937,
+ "eval_num_tokens": 1463356.0,
+ "eval_runtime": 98.6523,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 600
+ },
+ {
+ "entropy": 0.6257719069719314,
+ "epoch": 1.494571773220748,
+ "grad_norm": 0.48310142755508423,
+ "learning_rate": 7.176013751259459e-05,
+ "loss": 0.5713310241699219,
+ "mean_token_accuracy": 0.8305533796548843,
+ "num_tokens": 1512340.0,
+ "step": 620
+ },
+ {
+ "epoch": 1.494571773220748,
+ "eval_entropy": 0.6396401956510008,
+ "eval_loss": 0.5978795289993286,
+ "eval_mean_token_accuracy": 0.8273170630583603,
+ "eval_num_tokens": 1512340.0,
+ "eval_runtime": 98.4958,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 620
+ },
+ {
+ "entropy": 0.6162159442901611,
+ "epoch": 1.5428226779252112,
+ "grad_norm": 0.4500057101249695,
+ "learning_rate": 7.165127876018717e-05,
+ "loss": 0.565293264389038,
+ "mean_token_accuracy": 0.8311659932136536,
+ "num_tokens": 1560911.0,
+ "step": 640
+ },
+ {
+ "epoch": 1.5428226779252112,
+ "eval_entropy": 0.6381519051750054,
+ "eval_loss": 0.5963984131813049,
+ "eval_mean_token_accuracy": 0.8268409247478742,
+ "eval_num_tokens": 1560911.0,
+ "eval_runtime": 98.6803,
+ "eval_samples_per_second": 14.39,
+ "eval_steps_per_second": 1.804,
+ "step": 640
+ },
+ {
+ "entropy": 0.6088747605681419,
+ "epoch": 1.5910735826296745,
+ "grad_norm": 0.46954667568206787,
+ "learning_rate": 7.153237228458511e-05,
+ "loss": 0.5602522850036621,
+ "mean_token_accuracy": 0.8345131903886795,
+ "num_tokens": 1609409.0,
+ "step": 660
+ },
+ {
+ "epoch": 1.5910735826296745,
+ "eval_entropy": 0.6413098875056492,
+ "eval_loss": 0.5938988327980042,
+ "eval_mean_token_accuracy": 0.828771137119679,
+ "eval_num_tokens": 1609409.0,
+ "eval_runtime": 98.4951,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 660
+ },
+ {
+ "entropy": 0.6209868296980858,
+ "epoch": 1.6393244873341375,
+ "grad_norm": 0.3867878317832947,
+ "learning_rate": 7.140345173491446e-05,
+ "loss": 0.5663797378540039,
+ "mean_token_accuracy": 0.8352224260568619,
+ "num_tokens": 1657430.0,
+ "step": 680
+ },
+ {
+ "epoch": 1.6393244873341375,
+ "eval_entropy": 0.6182791748073664,
+ "eval_loss": 0.5919068455696106,
+ "eval_mean_token_accuracy": 0.8286421476455217,
+ "eval_num_tokens": 1657430.0,
+ "eval_runtime": 98.6211,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 680
+ },
+ {
+ "entropy": 0.6191923573613167,
+ "epoch": 1.6875753920386007,
+ "grad_norm": 0.40004226565361023,
+ "learning_rate": 7.12645535941658e-05,
+ "loss": 0.57517991065979,
+ "mean_token_accuracy": 0.8333286643028259,
+ "num_tokens": 1708058.0,
+ "step": 700
+ },
+ {
+ "epoch": 1.6875753920386007,
+ "eval_entropy": 0.6244879199212856,
+ "eval_loss": 0.5897895693778992,
+ "eval_mean_token_accuracy": 0.8293621252092083,
+ "eval_num_tokens": 1708058.0,
+ "eval_runtime": 98.4974,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 700
+ },
+ {
+ "entropy": 0.6199282392859459,
+ "epoch": 1.7358262967430638,
+ "grad_norm": 0.36575791239738464,
+ "learning_rate": 7.111571716886997e-05,
+ "loss": 0.5717356681823731,
+ "mean_token_accuracy": 0.8321028783917427,
+ "num_tokens": 1761262.0,
+ "step": 720
+ },
+ {
+ "epoch": 1.7358262967430638,
+ "eval_entropy": 0.6162440127201294,
+ "eval_loss": 0.5864423513412476,
+ "eval_mean_token_accuracy": 0.830874237786518,
+ "eval_num_tokens": 1761262.0,
+ "eval_runtime": 98.6109,
+ "eval_samples_per_second": 14.4,
+ "eval_steps_per_second": 1.805,
+ "step": 720
+ },
+ {
+ "entropy": 0.6127455830574036,
+ "epoch": 1.784077201447527,
+ "grad_norm": 0.4778473973274231,
+ "learning_rate": 7.095698457797475e-05,
+ "loss": 0.5664590835571289,
+ "mean_token_accuracy": 0.8324130088090896,
+ "num_tokens": 1812127.0,
+ "step": 740
+ },
+ {
+ "epoch": 1.784077201447527,
+ "eval_entropy": 0.6273092945640006,
+ "eval_loss": 0.5847376585006714,
+ "eval_mean_token_accuracy": 0.831051754817534,
+ "eval_num_tokens": 1812127.0,
+ "eval_runtime": 98.6513,
+ "eval_samples_per_second": 14.394,
+ "eval_steps_per_second": 1.804,
+ "step": 740
+ },
+ {
+ "entropy": 0.6141144648194313,
+ "epoch": 1.8323281061519903,
+ "grad_norm": 0.6467089056968689,
+ "learning_rate": 7.078840074092574e-05,
+ "loss": 0.5548617362976074,
+ "mean_token_accuracy": 0.8342376738786698,
+ "num_tokens": 1858403.0,
+ "step": 760
+ },
+ {
+ "epoch": 1.8323281061519903,
+ "eval_entropy": 0.5951976082967908,
+ "eval_loss": 0.5850171446800232,
+ "eval_mean_token_accuracy": 0.8303041870004675,
+ "eval_num_tokens": 1858403.0,
+ "eval_runtime": 98.6601,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 760
+ },
+ {
+ "entropy": 0.5961867064237595,
+ "epoch": 1.8805790108564535,
+ "grad_norm": 0.41280242800712585,
+ "learning_rate": 7.061001336495465e-05,
+ "loss": 0.5511267185211182,
+ "mean_token_accuracy": 0.8392465204000473,
+ "num_tokens": 1907361.0,
+ "step": 780
+ },
+ {
+ "epoch": 1.8805790108564535,
+ "eval_entropy": 0.6167017998320333,
+ "eval_loss": 0.5826791524887085,
+ "eval_mean_token_accuracy": 0.8309678050239434,
+ "eval_num_tokens": 1907361.0,
+ "eval_runtime": 98.6564,
+ "eval_samples_per_second": 14.393,
+ "eval_steps_per_second": 1.804,
+ "step": 780
+ },
+ {
+ "entropy": 0.6025628983974457,
+ "epoch": 1.9288299155609168,
+ "grad_norm": 0.4078947901725769,
+ "learning_rate": 7.042187293157881e-05,
+ "loss": 0.5552026271820069,
+ "mean_token_accuracy": 0.8359778299927711,
+ "num_tokens": 1956605.0,
+ "step": 800
+ },
+ {
+ "epoch": 1.9288299155609168,
+ "eval_entropy": 0.6047008484601974,
+ "eval_loss": 0.5811123847961426,
+ "eval_mean_token_accuracy": 0.8320730804057603,
+ "eval_num_tokens": 1956605.0,
+ "eval_runtime": 98.6033,
+ "eval_samples_per_second": 14.401,
+ "eval_steps_per_second": 1.805,
+ "step": 800
+ },
+ {
+ "entropy": 0.6041995719075203,
+ "epoch": 1.97708082026538,
+ "grad_norm": 0.543890118598938,
+ "learning_rate": 7.022403268231535e-05,
+ "loss": 0.554384422302246,
+ "mean_token_accuracy": 0.8380729153752327,
+ "num_tokens": 2003308.0,
+ "step": 820
+ },
+ {
+ "epoch": 1.97708082026538,
+ "eval_entropy": 0.6005127917179901,
+ "eval_loss": 0.5800131559371948,
+ "eval_mean_token_accuracy": 0.8313186547059691,
+ "eval_num_tokens": 2003308.0,
+ "eval_runtime": 98.4119,
+ "eval_samples_per_second": 14.429,
+ "eval_steps_per_second": 1.809,
+ "step": 820
+ },
+ {
+ "entropy": 0.567442889396961,
+ "epoch": 2.0241254523522314,
+ "grad_norm": 0.5015465021133423,
+ "learning_rate": 7.001654860361456e-05,
+ "loss": 0.5208251476287842,
+ "mean_token_accuracy": 0.8442784960453327,
+ "num_tokens": 2051331.0,
+ "step": 840
+ },
+ {
+ "epoch": 2.0241254523522314,
+ "eval_entropy": 0.5785037809878253,
+ "eval_loss": 0.5865591168403625,
+ "eval_mean_token_accuracy": 0.8307993857378371,
+ "eval_num_tokens": 2051331.0,
+ "eval_runtime": 98.5061,
+ "eval_samples_per_second": 14.415,
+ "eval_steps_per_second": 1.807,
+ "step": 840
+ },
+ {
+ "entropy": 0.5552012033760547,
+ "epoch": 2.0723763570566947,
+ "grad_norm": 0.5975943207740784,
+ "learning_rate": 6.979947941101646e-05,
+ "loss": 0.5064489841461182,
+ "mean_token_accuracy": 0.8468396887183189,
+ "num_tokens": 2101528.0,
+ "step": 860
+ },
+ {
+ "epoch": 2.0723763570566947,
+ "eval_entropy": 0.5844205239515626,
+ "eval_loss": 0.585284948348999,
+ "eval_mean_token_accuracy": 0.8309797065311604,
+ "eval_num_tokens": 2101528.0,
+ "eval_runtime": 98.4982,
+ "eval_samples_per_second": 14.417,
+ "eval_steps_per_second": 1.807,
+ "step": 860
+ },
+ {
+ "entropy": 0.5401638634502888,
+ "epoch": 2.120627261761158,
+ "grad_norm": 0.4194582998752594,
+ "learning_rate": 6.957288653253486e-05,
+ "loss": 0.49037814140319824,
+ "mean_token_accuracy": 0.8528496414422989,
+ "num_tokens": 2149550.0,
+ "step": 880
+ },
+ {
+ "epoch": 2.120627261761158,
+ "eval_entropy": 0.574507936835289,
+ "eval_loss": 0.5836747288703918,
+ "eval_mean_token_accuracy": 0.8325241643391298,
+ "eval_num_tokens": 2149550.0,
+ "eval_runtime": 98.5592,
+ "eval_samples_per_second": 14.408,
+ "eval_steps_per_second": 1.806,
+ "step": 880
+ },
+ {
+ "entropy": 0.5606910541653634,
+ "epoch": 2.168878166465621,
+ "grad_norm": 0.5395355820655823,
+ "learning_rate": 6.933683409127401e-05,
+ "loss": 0.5089571475982666,
+ "mean_token_accuracy": 0.8459660410881042,
+ "num_tokens": 2202162.0,
+ "step": 900
+ },
+ {
+ "epoch": 2.168878166465621,
+ "eval_entropy": 0.5779740023479033,
+ "eval_loss": 0.5816043615341187,
+ "eval_mean_token_accuracy": 0.8326318036304431,
+ "eval_num_tokens": 2202162.0,
+ "eval_runtime": 98.4386,
+ "eval_samples_per_second": 14.425,
+ "eval_steps_per_second": 1.808,
+ "step": 900
+ },
+ {
+ "entropy": 0.5380903735756875,
+ "epoch": 2.2171290711700844,
+ "grad_norm": 0.46588417887687683,
+ "learning_rate": 6.909138888728257e-05,
+ "loss": 0.4902662754058838,
+ "mean_token_accuracy": 0.8493925213813782,
+ "num_tokens": 2256539.0,
+ "step": 920
+ },
+ {
+ "epoch": 2.2171290711700844,
+ "eval_entropy": 0.570154261220707,
+ "eval_loss": 0.5827520489692688,
+ "eval_mean_token_accuracy": 0.8321914167216654,
+ "eval_num_tokens": 2256539.0,
+ "eval_runtime": 98.5674,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 920
+ },
+ {
+ "entropy": 0.558555257320404,
+ "epoch": 2.2653799758745476,
+ "grad_norm": 0.44576048851013184,
+ "learning_rate": 6.883662037864996e-05,
+ "loss": 0.5000699520111084,
+ "mean_token_accuracy": 0.8475695163011551,
+ "num_tokens": 2299698.0,
+ "step": 940
+ },
+ {
+ "epoch": 2.2653799758745476,
+ "eval_entropy": 0.5715417749761196,
+ "eval_loss": 0.5831062197685242,
+ "eval_mean_token_accuracy": 0.8327409899636601,
+ "eval_num_tokens": 2299698.0,
+ "eval_runtime": 98.568,
+ "eval_samples_per_second": 14.406,
+ "eval_steps_per_second": 1.806,
+ "step": 940
+ },
+ {
+ "entropy": 0.5582522705197335,
+ "epoch": 2.313630880579011,
+ "grad_norm": 0.38966405391693115,
+ "learning_rate": 6.857260066185045e-05,
+ "loss": 0.5132341861724854,
+ "mean_token_accuracy": 0.8466864034533501,
+ "num_tokens": 2347134.0,
+ "step": 960
+ },
+ {
+ "epoch": 2.313630880579011,
+ "eval_entropy": 0.5823918579334624,
+ "eval_loss": 0.5797588229179382,
+ "eval_mean_token_accuracy": 0.8326037566313583,
+ "eval_num_tokens": 2347134.0,
+ "eval_runtime": 98.6222,
+ "eval_samples_per_second": 14.398,
+ "eval_steps_per_second": 1.805,
+ "step": 960
+ },
+ {
+ "entropy": 0.5557353265583516,
+ "epoch": 2.361881785283474,
+ "grad_norm": 0.42300811409950256,
+ "learning_rate": 6.829940445134089e-05,
+ "loss": 0.5035375118255615,
+ "mean_token_accuracy": 0.8479988545179367,
+ "num_tokens": 2394568.0,
+ "step": 980
+ },
+ {
+ "epoch": 2.361881785283474,
+ "eval_entropy": 0.5542296482605881,
+ "eval_loss": 0.5800604820251465,
+ "eval_mean_token_accuracy": 0.8336404163515969,
+ "eval_num_tokens": 2394568.0,
+ "eval_runtime": 98.615,
+ "eval_samples_per_second": 14.399,
+ "eval_steps_per_second": 1.805,
+ "step": 980
+ }
+ ],
+ "logging_steps": 20,
+ "max_steps": 4150,
+ "num_input_tokens_seen": 0,
+ "num_train_epochs": 10,
+ "save_steps": 20,
+ "stateful_callbacks": {
+ "TrainerControl": {
+ "args": {
+ "should_epoch_stop": false,
+ "should_evaluate": false,
+ "should_log": false,
+ "should_save": true,
+ "should_training_stop": false
+ },
+ "attributes": {}
+ }
+ },
+ "total_flos": 4.97595589134336e+17,
+ "train_batch_size": 8,
+ "trial_name": null,
+ "trial_params": null
+}